Hugging Face's logo

Datasets:
EasonFan
/
AirCopBench

Modalities:
Image
Dataset card Data Studio Files
xet
Community
AirCopBench / test /Sim5_VQA_test.json
EasonFan's picture
EasonFan
Add files using upload-large-folder tool
c3e6555 verified
raw
history blame contribute delete
232 kB
 [
{
"sequence_frame": "4-40m-1629093666833511680.png",
"question_id": "MDMT_CA_UAV3_18",
"question_type": "Causal Assessment",
"question": "What is the primary cause of object detection challenges in this UAV3 image?",
"options": {
"A": "High image noise caused by sensor failure",
"B": "Occlusion by dense tree placement",
"C": "Low detail resolution from distant camera positioning",
"D": "Overexposure due to excessive sunlight"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a landscape with rocky terrain and dense trees. The trees are positioned close together, causing occlusion of underlying objects or potential points of interest, which can hinder accurate object detection.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629093666833511680-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629093666833511680-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629093666833511680-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629093666833511680-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629093666833511680-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624114615476825088.png",
"question_id": "MDMT_OC_UAV2_13",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV2's perspective?",
"options": {
"A": "21",
"B": "20",
"C": "18",
"D": "19"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624114615476825088-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624114615476825088-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624114615476825088-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624114615476825088-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624114615476825088-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629210769328715776.png",
"question_id": "MDMT_UA_UAV3_29",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV3 usable for target perception tasks?",
"options": {
"A": "Yes, highly usable",
"B": "Yes, barely usable",
"C": "Yes, partially usable",
"D": "No, not usable"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV3",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629210769328715776-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629210769328715776-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629210769328715776-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629210769328715776-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629210769328715776-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107791289598720.png",
"question_id": "MDMT_CA_UAV5_8",
"question_type": "Causal Assessment",
"question": "What is the main factor reducing visibility of objects in the lower center of the image?",
"options": {
"A": "Shadows cast by tall structures",
"B": "Overexposure caused by bright sunlight",
"C": "Blur due to a fast camera motion",
"D": "Occlusion by nearby trees"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a street scene captured by UAV5 from a high altitude, featuring residential buildings, roads, and vehicles. A notable perception quality issue arises in the lower center of the image, where shadows cast by tall structures significantly obscure details of objects, reducing visibility.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107791289598720-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107791289598720-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107791289598720-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107791289598720-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107791289598720-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629187328497839616.png",
"question_id": "MDMT_SD_UAV2_25",
"question_type": "Scene Description",
"question": "What is the relationship between the forested area and the surrounding roads in this scene?",
"options": {
"A": "The forested area is directly adjacent to a single road without any intersections.",
"B": "The forested area is surrounded by multiple roads forming a loop.",
"C": "The forested area is scattered between distant and disconnected roads.",
"D": "The forested area is isolated with no roads nearby."
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 50.0877808988764, 'y': 22.6123595505618, 'width': 0.9480337078651644, 'height': 2.1067415730337054, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 47.48068820224718, 'y': 22.47191011235956, 'width': 1.343047752808997, 'height': 2.6685393258426977, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 48.58672752808989, 'y': 27.52808988764045, 'width': 0.7900280898876346, 'height': 2.528089887640448, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 46.69066011235955, 'y': 27.808988764044944, 'width': 0.5530196629213435, 'height': 1.6853932584269664, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 45.42661516853933, 'y': 27.52808988764045, 'width': 0.7110252808988733, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 46.76966292134831, 'y': 32.162921348314605, 'width': 1.1060393258426942, 'height': 2.528089887640448, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 45.42661516853933, 'y': 35.1123595505618, 'width': 0.7900280898876346, 'height': 1.5449438202247165, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 46.611657303370784, 'y': 37.640449438202246, 'width': 1.264044943820224, 'height': 1.68539325842697, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 45.11060393258427, 'y': 21.910112359550563, 'width': 0.9480337078651644, 'height': 1.6853932584269629, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 46.29564606741573, 'y': 21.06741573033708, 'width': 0.9480337078651644, 'height': 1.1235955056179776, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 48.58672752808989, 'y': 19.101123595505616, 'width': 0.7900280898876346, 'height': 1.9662921348314626, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The scene shows a central forested area surrounded by roads that form a loop. A main highway is situated nearby, with visible traffic flow and additional residential and urban infrastructure in the background.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629187328497839616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629187328497839616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629187328497839616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629187328497839616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629187328497839616-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "Sim5_when2col_UAV4_46",
"question_type": "When to Collaborate",
"question": "Should UAV4 collaborate with other UAVs to address need for collaboration due to environmental factors across UAVs?",
"options": {
"A": "Yes, due to poor lighting conditions in multiple views",
"B": "No, lighting is adequate across all UAVs",
"C": "Yes, due to low image resolution in some views",
"D": "No, the environment is clear for all UAVs"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: 1 (Yes)",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629210769328715776.png",
"question_id": "MDMT_OG_UAV5_28",
"question_type": "Object Grounding",
"question": "Where is the gray car located relative to the blue car in this scene?",
"options": {
"A": "The gray car is ahead of the blue car in the same lane",
"B": "The gray car is behind the blue car but in a different lane",
"C": "The gray car is adjacent to the blue car in the neighboring lane",
"D": "The gray car is directly in front of the blue car in a parallel lane"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image captures a highway scene from UAV5 showing several cars occupying multiple lanes. The gray car is positioned adjacent to the blue car in the neighboring lane. Surrounding the roadway are areas with forested terrain and open green spaces.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629210769328715776-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629210769328715776-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629210769328715776-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629210769328715776-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629210769328715776-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_OG_UAV4_43",
"question_type": "Object Grounding",
"question": "What is the position of the large white building relative to the group of four pink apartment buildings in the scene?",
"options": {
"A": "The white building is to the left of the pink apartment buildings across a road",
"B": "The white building is directly behind the pink apartment buildings separated by green trees",
"C": "The white building is in front of the pink apartment buildings along the main road",
"D": "The white building is diagonally opposite to the pink apartment buildings across an intersection"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene includes a large white building positioned on the left side of a road with four pink apartment buildings grouped together on the right. Palm trees flank the road, and the area behind the pink apartments features additional greenery.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628412222245574656.png",
"question_id": "Sim5_what2col_UAV4_37",
"question_type": "What to Collaborate",
"question": "What specific object information should UAV4 collaborate on to improve vehicle tracking accuracy across views?",
"options": {
"A": "Details of the movement and position of vehicles near the corner intersections",
"B": "Information about the types of buildings near the road based on architectural patterns",
"C": "Color and visibility details of vehicles shown only in UAV4's view",
"D": "Presence of pedestrians across the green space in the foreground area"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV4 provides a detailed view of corner intersections with multiple vehicles visible, while other UAVs show broader city scenes with additional perspectives of traffic, buildings, and open spaces.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628412222245574656-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628412222245574656-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628412222245574656-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628412222245574656-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628412222245574656-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629187328497839616.png",
"question_id": "Sim5_who2col_UAV4_25",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV4 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV2",
"B": "None (no need for collaboration)",
"C": "UAV3",
"D": "UAV1"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV2",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629187328497839616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629187328497839616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629187328497839616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629187328497839616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629187328497839616-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107811717034496.png",
"question_id": "MDMT_SC_9",
"question_type": "Scene Comparison",
"question": "Which UAV perspective provides better visibility of the residential area layout?",
"options": {
"A": "UAV1 provides the clearest view of the residential area with distinct details like houses and green spaces.",
"B": "UAV2 provides the clearest view of the residential area, showing houses and streets more prominently.",
"C": "UAV3 provides the best visibility of the residential area due to its higher altitude perspective.",
"D": "UAV4 provides the clearest view by emphasizing the overall grid layout of the residential area."
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Poor (2/5) - Blurry, noisy, color and brightness problems.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 39.50140449438202, 'y': 2.8089887640449436, 'width': 1.264044943820224, 'height': 2.1067415730337085, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 56.09199438202247, 'y': 2.9494382022471908, 'width': 1.5800561797752763, 'height': 2.52808988764045, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: None; Collaboration reasons: Current image misses the target due to out-of-view, but other views may capture it.",
"image_description": "UAV1 shows a detailed view of a residential neighborhood with clear visibility of houses, yards, and a central intersection. UAV2 focuses on urban areas with dense buildings and roads, UAV3 shows commercial structures, and UAV4 highlights buildings and green spaces but less of the residential layout.",
"missing_models": -12,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107811717034496-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107811717034496-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107811717034496-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107811717034496-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107811717034496-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171377571127296.png",
"question_id": "MDMT_OR_UAV1_16",
"question_type": "Object Recognition",
"question": "What type of prominent non-vehicle object is visible near the center of the image?",
"options": {
"A": "A pedestrian",
"B": "A cyclist",
"C": "A bus stop structure",
"D": "A streetlight pole"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene depicts an intersection with several vehicles on a multi-lane road. Near the center of the image, there is a large modern structure resembling a bus stop or station, with several smaller building-like structures nearby.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171377571127296-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171377571127296-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171377571127296-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171377571127296-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171377571127296-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "MDMT_OC_UAV2_31",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV2's perspective?",
"options": {
"A": "59",
"B": "56",
"C": "53",
"D": "54"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629204982295740416.png",
"question_id": "Sim5_who2col_UAV3_28",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV3 collaborate with for collaboration partner for specific object data across UAVs?",
"options": {
"A": "UAV2",
"B": "None (no need for collaboration)",
"C": "UAV5",
"D": "UAV4"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV2",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629204982295740416-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629204982295740416-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629204982295740416-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629204982295740416-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629204982295740416-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628409917747699200.png",
"question_id": "MDMT_OC_UAV1_35",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV1's perspective?",
"options": {
"A": "39",
"B": "40",
"C": "37",
"D": "42"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628409917747699200-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628409917747699200-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628409917747699200-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628409917747699200-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628409917747699200-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628248254973832704.png",
"question_id": "MDMT_SD_UAV2_34",
"question_type": "Scene Description",
"question": "What type of setting is depicted in the image based on the layout of the buildings, roads, and vegetation?",
"options": {
"A": "A densely packed suburban neighborhood with medium-sized buildings and scattered vegetation",
"B": "A rural area with vast open fields and sparse buildings",
"C": "An urban area featuring a mix of tall skyscrapers, structured roads, and patches of greenery",
"D": "An industrial zone with factories and warehouses surrounded by roads"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 63.36025280898876, 'y': 11.938202247191011, 'width': 1.1060393258427013, 'height': 1.5449438202247183, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 65.33532303370787, 'y': 10.955056179775282, 'width': 1.3430477528089853, 'height': 1.4044943820224702, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 63.36025280898876, 'y': 8.98876404494382, 'width': 1.0270365168539257, 'height': 2.3876404494382015, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 65.2563202247191, 'y': 8.426966292134832, 'width': 1.3430477528089853, 'height': 1.5449438202247183, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 67.70540730337078, 'y': 8.98876404494382, 'width': 0.9480337078651786, 'height': 1.5449438202247183, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 68.89044943820224, 'y': 6.03932584269663, 'width': 0.7900280898876559, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 67.15238764044943, 'y': 7.02247191011236, 'width': 0.8690308988764031, 'height': 1.404494382022472, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 65.2563202247191, 'y': 6.460674157303371, 'width': 0.9480337078651644, 'height': 1.5449438202247183, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 68.02141853932584, 'y': 4.915730337078652, 'width': 0.9480337078651786, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 66.59936797752809, 'y': 3.089887640449438, 'width': 1.2640449438202381, 'height': 2.1067415730337076, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The scene shows a structured urban area with a mix of tall buildings, including a cylindrical skyscraper and multi-tiered structures, wide roads, and patches of greenery and trees separating the buildings and roads.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628248254973832704-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628248254973832704-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628248254973832704-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628248254973832704-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628248254973832704-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629093666833511680.png",
"question_id": "MDMT_OG_UAV5_17",
"question_type": "Object Grounding",
"question": "What is the position of the smaller house relative to the larger house near the water edge?",
"options": {
"A": "The smaller house is located uphill and further from the water compared to the larger house.",
"B": "The smaller house is positioned downhill and closer to the water compared to the larger house.",
"C": "The smaller house is beside the larger house, equidistant from the water.",
"D": "The smaller house is further away from the trees compared to the larger house."
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a coastal scene with two houses near the water edge. The larger house is located uphill, while the smaller house is positioned downhill closer to the water. The landscape includes trees, rocks, and a winding road in the background.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629093666833511680-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629093666833511680-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629093666833511680-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629093666833511680-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629093666833511680-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105318731299328.png",
"question_id": "Sim5_why2col_UAV2_6",
"question_type": "Why to Collaborate",
"question": "What is the main reason UAV2 should collaborate with other UAVs in this scenario?",
"options": {
"A": "To compensate for the small and distant target size visible in UAV2's view",
"B": "To synchronize overlapping data from identical regions",
"C": "To improve visibility of shadowed objects in UAV2's perspective",
"D": "To generate a shared panoramic image across multiple perspectives"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The image from UAV2 shows a road scene where distant vehicles appear too small for clear recognition. However, other UAV perspectives show these vehicles larger and more visible, providing complementary details.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105318731299328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105318731299328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105318731299328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105318731299328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105318731299328-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628409917747699200.png",
"question_id": "MDMT_QA_UAV1_35",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV1?",
"options": {
"A": "2",
"B": "1",
"C": "3",
"D": "5"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV1",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628409917747699200-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628409917747699200-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628409917747699200-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628409917747699200-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628409917747699200-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623937902840830464.png",
"question_id": "MDMT_OC_UAV5_4",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV5's perspective?",
"options": {
"A": "7",
"B": "12",
"C": "9",
"D": "8"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623937902840830464-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623937902840830464-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623937902840830464-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623937902840830464-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623937902840830464-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105298882875904.png",
"question_id": "MDMT_QA_UAV2_5",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV2?",
"options": {
"A": "5",
"B": "3",
"C": "4",
"D": "1"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105298882875904-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105298882875904-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105298882875904-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105298882875904-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105298882875904-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628497853920271360.png",
"question_id": "MDMT_SD_UAV4_45",
"question_type": "Scene Description",
"question": "What is the primary environmental context depicted in this scene?",
"options": {
"A": "A densely populated urban area with skyscrapers",
"B": "A suburban area with buildings and open spaces",
"C": "A remote rural area dominated by farmland",
"D": "A forested region with no human structures"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Poor (2/5) - Blurry, noisy, color and brightness problems.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 79.79283707865167, 'y': 44.10112359550562, 'width': 3.160112359550567, 'height': 2.668539325842694, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 51.1938202247191, 'y': 30.89887640449438, 'width': 1.896067415730343, 'height': 2.106741573033709, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 43.688553370786515, 'y': 14.747191011235955, 'width': 1.1850421348314626, 'height': 1.54494382022472, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 43.530547752808985, 'y': 17.134831460674157, 'width': 1.1060393258427013, 'height': 1.2640449438202275, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "The scene shows a combination of urban and suburban features, including rows of multi-story buildings adjacent to a road, smaller individual houses, open grassy spaces, and other structures such as silos and a church.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628497853920271360-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628497853920271360-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628497853920271360-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628497853920271360-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628497853920271360-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_SD_UAV4_46",
"question_type": "Scene Description",
"question": "What type of setting does the environmental layout in this scene primarily represent?",
"options": {
"A": "A suburban area with a mix of residential buildings and green spaces",
"B": "A dense urban area with tall skyscrapers dominating the landscape",
"C": "A rural area with open fields and sporadic buildings",
"D": "An industrial area with warehouses and factories"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 70.23349719101122, 'y': 10.53370786516854, 'width': 1.3430477528089995, 'height': 1.264044943820224, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 70.78651685393258, 'y': 8.146067415730338, 'width': 1.264044943820224, 'height': 1.264044943820224, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 70.39150280898876, 'y': 9.410112359550562, 'width': 1.8170646067415677, 'height': 0.8426966292134831, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The image shows a suburban area with rows of residential buildings, small houses, paved roads, and scattered green spaces. Palm trees line the roads, and the buildings are arranged in an orderly manner, emphasizing a planned urban setting rather than a dense skyline or wide-open rural environment.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629211736846056704.png",
"question_id": "Sim5_who2col_UAV4_30",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV4 collaborate with for collaboration partner for specific object data across UAVs?",
"options": {
"A": "UAV3",
"B": "None (no need for collaboration)",
"C": "UAV1",
"D": "UAV5"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV3",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629211736846056704-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629211736846056704-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629211736846056704-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629211736846056704-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629211736846056704-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628423268493889280.png",
"question_id": "Sim5_why2col_UAV4_40",
"question_type": "Why to Collaborate",
"question": "Why is it necessary for UAV4 to collaborate with other UAVs based on the image and annotation data?",
"options": {
"A": "To enhance visibility of small or distant targets that are clearer in other views",
"B": "To ensure synchronization between all UAVs in the area",
"C": "To compensate for low altitude limitations that reduce field of view",
"D": "To reduce redundant image data captured by all UAVs"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The target in UAV4's view appears too small or distant for clear identification, but is larger and more visible in images from other UAV perspectives.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628423268493889280-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628423268493889280-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628423268493889280-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628423268493889280-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628423268493889280-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "Sim5_who2col_UAV5_31",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV5 collaborate with for collaboration partner for specific object data across UAVs?",
"options": {
"A": "UAV3",
"B": "None (no need for collaboration)",
"C": "UAV4",
"D": "UAV2"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV3",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "MDMT_SD_UAV3_31",
"question_type": "Scene Description",
"question": "What is the relationship between the vehicles and the road environment in this scene?",
"options": {
"A": "Vehicles are parked along the road edges, with wide space between lanes",
"B": "Vehicles are moving in dense traffic along a curved multi-lane road",
"C": "Vehicles are scattered on the road without following lanes",
"D": "Vehicles are stationary on a single-lane road surrounded tightly by buildings"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Poor (2/5) - Blurry, noisy, color and brightness problems.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 40.60744382022472, 'y': 11.095505617977528, 'width': 1.1060393258427013, 'height': 1.9662921348314626, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 45.42661516853933, 'y': 11.51685393258427, 'width': 1.501053370786508, 'height': 1.404494382022472, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 48.04579893440832, 'y': 7.690164496519157, 'width': 1.2932557561800877, 'height': 1.4270408344056174, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 49.11607956021252, 'y': 6.500963801181143, 'width': 1.2040657040297447, 'height': 1.1892006953380134, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 50.00798008171604, 'y': 5.073922966775527, 'width': 1.2040657040297305, 'height': 1.2684807416938817, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 51.47961594219683, 'y': 5.708163337622468, 'width': 1.2486607301049233, 'height': 0.9513605562704104, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 50.94447562929473, 'y': 6.976644079316349, 'width': 1.5162308865559595, 'height': 0.9513605562704113, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 53.21882195912868, 'y': 4.201842456860983, 'width': 1.0702806258042088, 'height': 1.2684807416938817, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 57.232374305894474, 'y': 3.7261621787257773, 'width': 1.3824458083304307, 'height': 0.7928004635586761, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 61.3797117308858, 'y': 4.201842456860983, 'width': 1.248660730104909, 'height': 1.6648809734732195, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 62.98513266959211, 'y': 2.854081668811234, 'width': 1.7392060169318526, 'height': 1.347760788049749, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 65.9284043905537, 'y': 4.043282364149247, 'width': 1.2040657040297447, 'height': 1.2684807416938817, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 67.35544522495931, 'y': 2.933361715167101, 'width': 1.2040657040297447, 'height': 1.347760788049749, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 70.03114678946984, 'y': 2.774801622455366, 'width': 1.4270408344056307, 'height': 1.8234410661849556, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 74.8474096055888, 'y': 3.7261621787257773, 'width': 0.9810905736538729, 'height': 0.7135204172028078, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 77.38932609187381, 'y': 3.7261621787257773, 'width': 1.0256855997290444, 'height': 0.8720805099145443, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 76.89878080504687, 'y': 4.915362874063792, 'width': 1.4716358604808022, 'height': 0.7928004635586756, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 77.56770619617451, 'y': 6.500963801181143, 'width': 1.5608259126311452, 'height': 0.9513605562704113, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 79.88664755208363, 'y': 4.915362874063792, 'width': 1.2486607301049304, 'height': 0.9513605562704104, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 80.6893580214368, 'y': 3.4090419933023073, 'width': 1.5162308865559595, 'height': 0.9513605562704104, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 81.0461182300382, 'y': 5.470323198554865, 'width': 1.4270408344056165, 'height': 1.1892006953380134, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV3; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The scene shows a curved and multi-lane road beside a waterfront, with vehicles moving in traffic. Surrounding the area are scattered trees and rocky terrain.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629109485061830400.png",
"question_id": "MDMT_OC_UAV3_23",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV3's perspective?",
"options": {
"A": "26",
"B": "28",
"C": "30",
"D": "29"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629109485061830400-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629109485061830400-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629109485061830400-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629109485061830400-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629109485061830400-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107791289598720.png",
"question_id": "Sim5_what2col_UAV2_8",
"question_type": "What to Collaborate",
"question": "What specific object information should UAV2 share with other UAVs to enhance scene understanding across multi-UAV views?",
"options": {
"A": "Positions and movements of vehicles passing under the elevated railway from UAV2's perspective",
"B": "Spatial layout and types of buildings near the railway area visible in UAV2’s view",
"C": "Details of pedestrian activity near the railway visible only to UAV2",
"D": "Traffic light locations and status at intersections visible to UAV2"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV2 shows an elevated railway with vehicles passing underneath, which is complemented by other UAVs providing views of different urban areas including residential zones, intersections, and a roundabout.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107791289598720-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107791289598720-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107791289598720-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107791289598720-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107791289598720-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "MDMT_OP_31",
"question_type": "Observing Posture",
"question": "Which UAV perspective provides a clearer understanding of the spatial layout of the apartment complex relative to nearby roads?",
"options": {
"A": "UAV1 provides a clearer view due to its broader angle showing surrounding areas.",
"B": "UAV2 provides a clearer view due to its close focus directly above the apartment complex.",
"C": "Both UAVs equally reveal the spatial layout of the apartment complex and nearby roads.",
"D": "Neither UAV provides a clear understanding of the spatial layout."
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Poor (2/5) - Blurry, noisy, color and brightness problems.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 40.60744382022472, 'y': 11.095505617977528, 'width': 1.1060393258427013, 'height': 1.9662921348314626, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 45.42661516853933, 'y': 11.51685393258427, 'width': 1.501053370786508, 'height': 1.404494382022472, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 48.04579893440832, 'y': 7.690164496519157, 'width': 1.2932557561800877, 'height': 1.4270408344056174, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 49.11607956021252, 'y': 6.500963801181143, 'width': 1.2040657040297447, 'height': 1.1892006953380134, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 50.00798008171604, 'y': 5.073922966775527, 'width': 1.2040657040297305, 'height': 1.2684807416938817, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 51.47961594219683, 'y': 5.708163337622468, 'width': 1.2486607301049233, 'height': 0.9513605562704104, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 50.94447562929473, 'y': 6.976644079316349, 'width': 1.5162308865559595, 'height': 0.9513605562704113, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 53.21882195912868, 'y': 4.201842456860983, 'width': 1.0702806258042088, 'height': 1.2684807416938817, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 57.232374305894474, 'y': 3.7261621787257773, 'width': 1.3824458083304307, 'height': 0.7928004635586761, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 61.3797117308858, 'y': 4.201842456860983, 'width': 1.248660730104909, 'height': 1.6648809734732195, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 62.98513266959211, 'y': 2.854081668811234, 'width': 1.7392060169318526, 'height': 1.347760788049749, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 65.9284043905537, 'y': 4.043282364149247, 'width': 1.2040657040297447, 'height': 1.2684807416938817, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 67.35544522495931, 'y': 2.933361715167101, 'width': 1.2040657040297447, 'height': 1.347760788049749, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 70.03114678946984, 'y': 2.774801622455366, 'width': 1.4270408344056307, 'height': 1.8234410661849556, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 74.8474096055888, 'y': 3.7261621787257773, 'width': 0.9810905736538729, 'height': 0.7135204172028078, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 77.38932609187381, 'y': 3.7261621787257773, 'width': 1.0256855997290444, 'height': 0.8720805099145443, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 76.89878080504687, 'y': 4.915362874063792, 'width': 1.4716358604808022, 'height': 0.7928004635586756, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 77.56770619617451, 'y': 6.500963801181143, 'width': 1.5608259126311452, 'height': 0.9513605562704113, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 79.88664755208363, 'y': 4.915362874063792, 'width': 1.2486607301049304, 'height': 0.9513605562704104, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 80.6893580214368, 'y': 3.4090419933023073, 'width': 1.5162308865559595, 'height': 0.9513605562704104, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 81.0461182300382, 'y': 5.470323198554865, 'width': 1.4270408344056165, 'height': 1.1892006953380134, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV3; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "UAV1 shows a wide view encompassing the apartment complex, surrounding residential homes, and nearby roads; UAV2 focuses closer on the apartment complex but limits visibility of surrounding roads.",
"missing_models": -12,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623936250982352640.png",
"question_id": "MDMT_OR_UAV1_3",
"question_type": "Object Recognition",
"question": "What type of non-vehicle object is most prominent in the scene shown from UAV1?",
"options": {
"A": "Residential buildings",
"B": "Tall office complexes",
"C": "Public park structures",
"D": "Industrial warehouses"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The UAV1 image shows a suburban scene with rows of residential houses, a fenced grassy area, and streets with multiple vehicles. Trees and fences separate the residential properties from the road.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623936250982352640-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623936250982352640-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623936250982352640-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623936250982352640-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623936250982352640-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629095239127567616.png",
"question_id": "MDMT_OG_UAV2_18",
"question_type": "Object Grounding",
"question": "What is the spatial position of the red truck relative to the building with flat roofs in the parking area?",
"options": {
"A": "The red truck is parked behind the building with flat roofs near the corner",
"B": "The red truck is positioned to the right side of the building with flat roofs, in the adjacent road",
"C": "The red truck is parked on the left side of the building with flat roofs, near the parking spaces",
"D": "The red truck is directly in front of the building with flat roofs, closest to the parking lot entrance"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows an aerial view of a parking area bordered by buildings. A red truck is located near the left side of a flat-roofed building adjacent to parking spaces. The scene includes residential buildings, multi-story structures, and parking arrangements with several parked cars and small trucks.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629095239127567616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629095239127567616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629095239127567616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629095239127567616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629095239127567616-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171352874600448.png",
"question_id": "MDMT_OR_UAV5_15",
"question_type": "Object Recognition",
"question": "What type of vehicle is visible on the road near the middle portion of the image?",
"options": {
"A": "A black sedan",
"B": "A red sports car",
"C": "A gray minivan",
"D": "A white pickup truck"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image captures a UAV perspective over a neighborhood with roads, trees, and buildings. In the middle of the scene, a black sedan is driving along a road lined with trees. Residential houses are visible to the right, and a concrete plaza with trees appears on the left.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171352874600448-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171352874600448-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171352874600448-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171352874600448-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171352874600448-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629091952438388224.png",
"question_id": "Sim5_who2col_UAV3_17",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV3 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV4",
"B": "None (no need for collaboration)",
"C": "UAV1",
"D": "UAV5"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV4",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629091952438388224-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629091952438388224-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629091952438388224-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629091952438388224-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629091952438388224-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628426933546515456.png",
"question_id": "MDMT_OR_UAV3_41",
"question_type": "Object Recognition",
"question": "What type of vehicle is visible on the road in this scene?",
"options": {
"A": "A black sedan",
"B": "A red SUV",
"C": "A white van",
"D": "A gray truck"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene depicts a forested area with a curved road running alongside a building and a grassy hill. Several vehicles, including a red SUV, are visible on the road.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628426933546515456-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628426933546515456-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628426933546515456-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628426933546515456-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628426933546515456-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629211736846056704.png",
"question_id": "Sim5_who2col_UAV2_30",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV2 collaborate with for collaboration partner for specific object data across UAVs?",
"options": {
"A": "UAV3",
"B": "None (no need for collaboration)",
"C": "UAV4",
"D": "UAV1"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV3",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629211736846056704-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629211736846056704-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629211736846056704-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629211736846056704-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629211736846056704-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628409917747699200.png",
"question_id": "Sim5_why2col_UAV5_35",
"question_type": "Why to Collaborate",
"question": "Why should UAV5 collaborate with other UAVs in this environment?",
"options": {
"A": "To overcome occlusion issues affecting the visibility of objects",
"B": "To ensure faster data transfer between UAVs",
"C": "To reduce overlaps in object detection across multiple views",
"D": "To improve positioning accuracy of UAV5 based on other UAV data"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "The current image from UAV5 suffers from object occlusion, while images from other UAVs provide complementary views with fewer obstructions.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628409917747699200-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628409917747699200-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628409917747699200-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628409917747699200-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628409917747699200-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628426933546515456.png",
"question_id": "Sim5_why2col_UAV4_41",
"question_type": "Why to Collaborate",
"question": "Why should UAV4 collaborate with other UAVs in this scene?",
"options": {
"A": "To resolve occlusion caused by dense trees blocking building views",
"B": "To obtain higher resolution images of the road infrastructure",
"C": "To provide detailed information on distant regions outside UAV4's field of view",
"D": "To track vehicle movement from multiple perspectives for improved motion analysis"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot Fallback)",
"annotation_info": "",
"image_description": "UAV4's view is obstructed by dense trees that block portions of the buildings and surrounding areas, while other UAVs provide complementary perspectives with better visibility.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628426933546515456-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628426933546515456-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628426933546515456-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628426933546515456-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628426933546515456-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_OG_UAV1_45",
"question_type": "Object Grounding",
"question": "Which object is directly across the intersection from the row of apartment buildings with four towers?",
"options": {
"A": "The large square building with surrounding trees",
"B": "The small set of residential houses near the forest",
"C": "The parking lot with vehicles",
"D": "The tall apartment complex with rectangular courtyards"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a residential area with a row of apartment buildings consisting of four towers situated near an intersection. Across the intersection is a large square building surrounded by trees, while nearby are other residential structures and parking areas with vehicles visible.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629096982293769472.png",
"question_id": "Sim5_who2col_UAV1_20",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV1 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV4",
"B": "None (no need for collaboration)",
"C": "UAV3",
"D": "UAV2"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV1",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV4",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629096982293769472-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629096982293769472-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629096982293769472-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629096982293769472-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629096982293769472-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "Sim5_who2col_UAV2_31",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV2 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV3",
"B": "None (no need for collaboration)",
"C": "UAV1",
"D": "UAV4"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV3",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623936157944367872.png",
"question_id": "Sim5_what2col_UAV4_1",
"question_type": "What to Collaborate",
"question": "What specific object information should UAV4 collaborate on with other UAVs to enhance multi-view perception?",
"options": {
"A": "Precise positions and movements of the vehicles near the roundabout for accurate traffic analysis",
"B": "Detailed appearance and orientation of the trees along the middle road for environmental modeling",
"C": "Locations and identities of stationary objects near the residential area for scene reconstruction",
"D": "Visibility and shadow effects of buildings in all UAV views for lighting condition estimation"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV4 shows vehicles moving through a straight road surrounded by residential areas and trees, while the other UAV views highlight traffic near the roundabout, building layouts, and roadside objects.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623936157944367872-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623936157944367872-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623936157944367872-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623936157944367872-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623936157944367872-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105298882875904.png",
"question_id": "MDMT_OR_UAV2_5",
"question_type": "Object Recognition",
"question": "What type of non-vehicle object is prominently visible near the buildings?",
"options": {
"A": "Streetlights",
"B": "Bicycles",
"C": "Trash bins",
"D": "Pedestrians"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a city scene with several buildings, vehicles on the road, and prominently visible streetlights near the buildings and on the sidewalks.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105298882875904-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105298882875904-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105298882875904-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105298882875904-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105298882875904-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628412914992445440.png",
"question_id": "MDMT_QA_UAV2_38",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV2?",
"options": {
"A": "5",
"B": "2",
"C": "4",
"D": "1"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628412914992445440-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628412914992445440-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628412914992445440-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628412914992445440-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628412914992445440-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629095239127567616.png",
"question_id": "MDMT_OC_UAV1_19",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV1's perspective?",
"options": {
"A": "20",
"B": "17",
"C": "18",
"D": "16"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629095239127567616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629095239127567616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629095239127567616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629095239127567616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629095239127567616-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_OR_UAV3_46",
"question_type": "Object Recognition",
"question": "What type of vehicle is most prominently visible near the intersection in the lower-left area of the scene?",
"options": {
"A": "A blue sedan",
"B": "A green pickup truck",
"C": "A red sports car",
"D": "A yellow bus"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene shows an aerial view of a residential area, including rows of mid-rise apartment buildings, individual houses, and streets with scattered vehicles. The lower-left area of the image includes an intersection where a blue sedan is prominently visible.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624106452866560512.png",
"question_id": "MDMT_OG_UAV4_7",
"question_type": "Object Grounding",
"question": "What is the position of the modern fountain relative to the cluster of four identical houses?",
"options": {
"A": "The fountain is to the left of the houses on the main road in the traffic circle",
"B": "The fountain is directly behind the houses in a residential parking lot",
"C": "The fountain is located in front of the houses within a square park area",
"D": "The fountain is to the right of the houses near a multi-lane intersection"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a modern fountain located at the center of a traffic circle. Nearby, there are four identical houses positioned along a curved street to the right of the fountain. The area includes trees, roads, and a residential neighborhood surrounding the scene.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624106452866560512-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624106452866560512-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624106452866560512-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624106452866560512-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624106452866560512-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628426933546515456.png",
"question_id": "MDMT_CA_UAV3_41",
"question_type": "Causal Assessment",
"question": "What is the primary factor affecting object detection in this image?",
"options": {
"A": "Overexposure caused by excessive sunlight",
"B": "Occlusion by foreground trees",
"C": "Shadows obstructing object visibility",
"D": "Blurring due to motion of the UAV"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows multiple objects partially obscured by foreground trees. The dense foliage and irregular placement of trees block portions of the roadway, making it challenging for object detection systems to clearly identify vehicles and other elements within the scene.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628426933546515456-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628426933546515456-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628426933546515456-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628426933546515456-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628426933546515456-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629187328497839616.png",
"question_id": "MDMT_OG_UAV1_25",
"question_type": "Object Grounding",
"question": "Which vehicle is positioned closest to the intersection located at the center of the scene?",
"options": {
"A": "A black car moving on the upper-left road",
"B": "A yellow truck near the center of the intersection",
"C": "A red car traveling on the upper-right road",
"D": "A purple car on the lower road beneath the intersection"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image depicts a large intersection surrounded by forested terrain. Several vehicles are visible across multiple lanes, including a black car on the upper-left road, red and yellow vehicles near the intersection, and a purple car on the lower road.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629187328497839616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629187328497839616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629187328497839616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629187328497839616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629187328497839616-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629201000237519360.png",
"question_id": "MDMT_OM_UAV5_27",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the group of pine trees near the cliff edge in UAV5?",
"options": {
"A": "The group of pine trees near the fenced residential area in UAV3",
"B": "The group of pine trees seen beyond the red-roofed house in UAV4",
"C": "The group of pine trees beside the large roadway curve in UAV2",
"D": "The group of pine trees directly overlooking the riverbank in UAV1"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV5 shows a group of pine trees near the cliff edge, overlooking the river. UAV1 shows the same group of trees directly overlooking the riverbank at a closer perspective. UAV2, UAV3, and UAV4 display other clusters of pine trees near residential zones, roads, or landmarks.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629201000237519360-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629201000237519360-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629201000237519360-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629201000237519360-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629201000237519360-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623936250982352640.png",
"question_id": "MDMT_OR_UAV3_3",
"question_type": "Object Recognition",
"question": "What type of vehicle is parked closest to the stone wall on the left side of the road?",
"options": {
"A": "A white sedan",
"B": "A black SUV",
"C": "A beige minivan",
"D": "A red pickup truck"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The aerial scene depicts a residential area with multiple vehicles driving along and parked near a road. A stone wall surrounds a prominent multi-story building. Several trees align the road on the right, and homes with red and brown rooftops are visible on the other side. An SUV is parked closest to the stone wall on the left side of the road.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623936250982352640-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623936250982352640-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623936250982352640-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623936250982352640-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623936250982352640-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628422473693368320.png",
"question_id": "MDMT_OC_UAV2_39",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV2's perspective?",
"options": {
"A": "34",
"B": "36",
"C": "32",
"D": "33"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628422473693368320-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628422473693368320-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628422473693368320-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628422473693368320-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628422473693368320-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628412222245574656.png",
"question_id": "MDMT_OM_UAV1_37",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the stone structure located near the curved road in UAV3?",
"options": {
"A": "The stone structure now seen from the front with trees in the background in UAV4",
"B": "The stone structure located closer to a row of identical houses in UAV1",
"C": "The stone structure now seen with palm trees surrounding it in UAV3",
"D": "The stone structure located next to cylindrical skyscrapers in UAV2"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV1 shows various apartment complexes and high-rise buildings; UAV2 highlights tall cylindrical skyscrapers near a white monument; UAV3 features a stone structure near a curved road surrounded by palm trees, visible from a different angle in UAV4 along a larger open field.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628412222245574656-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628412222245574656-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628412222245574656-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628412222245574656-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628412222245574656-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628410764252718848.png",
"question_id": "MDMT_UA_UAV5_36",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV5 usable for target perception tasks?",
"options": {
"A": "No, not usable",
"B": "Yes, highly usable",
"C": "Yes, barely usable",
"D": "Yes, partially usable"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628410764252718848-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628410764252718848-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628410764252718848-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628410764252718848-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628410764252718848-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105298882875904.png",
"question_id": "MDMT_SC_5",
"question_type": "Scene Comparison",
"question": "Which UAV perspective offers a clearer view of the elevated railway structure's surroundings?",
"options": {
"A": "UAV1 shows the elevated railway with adjacent buildings and vehicles",
"B": "UAV3 provides the clearest view of the elevated railway and its adjacent greenery",
"C": "UAV4 provides the clearest view of the elevated railway and traffic concentration nearby",
"D": "UAV5 does not show the elevated railway at all"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 0.15800561797752807, 'y': 50.28089887640449, 'width': 2.291081460674157, 'height': 5.337078651685395, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: None; Collaboration reasons: Current image misses the target due to out-of-view, but other views may capture it.",
"image_description": "UAV1 captures the elevated railway structure in an urban area with buildings and vehicles nearby. UAV3 shifts focus to a road adjacent to greenery, but the elevated railway is visible further away. UAV4 provides a detailed view of the railway surrounded by dense traffic and urban infrastructure. UAV5 focuses on rural-like areas without showing the railway.",
"missing_models": -9,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105298882875904-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105298882875904-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105298882875904-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105298882875904-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105298882875904-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629096982293769472.png",
"question_id": "Sim5_who2col_UAV5_20",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV5 collaborate with for collaboration partner for specific object data across UAVs?",
"options": {
"A": "UAV4",
"B": "None (no need for collaboration)",
"C": "UAV2",
"D": "UAV3"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV4",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629096982293769472-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629096982293769472-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629096982293769472-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629096982293769472-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629096982293769472-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624106452866560512.png",
"question_id": "MDMT_UA_UAV1_7",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV1 usable for target perception tasks?",
"options": {
"A": "Yes, highly usable",
"B": "Yes, barely usable",
"C": "Yes, partially usable",
"D": "No, not usable"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV1",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624106452866560512-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624106452866560512-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624106452866560512-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624106452866560512-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624106452866560512-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624114615476825088.png",
"question_id": "MDMT_OG_UAV2_12",
"question_type": "Object Grounding",
"question": "What is the spatial position of the parked black car relative to the nearby red car in this scene?",
"options": {
"A": "The black car is parked in a lane next to the red car",
"B": "The black car is ahead of the red car in the same lane",
"C": "The black car is parked in the same lane behind the red car",
"D": "The black car is on the opposite side of the road from the red car"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene depicts a road with multiple lanes. A black car is parked in the same lane behind a red car. There are several buildings and sidewalks visible in the surrounding area.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624114615476825088-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624114615476825088-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624114615476825088-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624114615476825088-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624114615476825088-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623937902840830464.png",
"question_id": "MDMT_UA_UAV2_4",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV2 usable for target perception tasks?",
"options": {
"A": "Yes, highly usable",
"B": "No, not usable",
"C": "Yes, partially usable",
"D": "Yes, barely usable"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623937902840830464-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623937902840830464-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623937902840830464-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623937902840830464-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623937902840830464-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107811717034496.png",
"question_id": "MDMT_OR_UAV5_9",
"question_type": "Object Recognition",
"question": "What type of non-vehicle object is most prominently visible along the sidewalk in the scene?",
"options": {
"A": "Street lamps",
"B": "Pedestrians",
"C": "Bicycles",
"D": "Trash bins"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image displays an urban road flanked by buildings with visible street lamps along the sidewalks. Multiple vehicles are seen driving on the road, and no pedestrians or cyclists are present.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107811717034496-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107811717034496-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107811717034496-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107811717034496-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107811717034496-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628412914992445440.png",
"question_id": "Sim5_what2col_UAV2_38",
"question_type": "What to Collaborate",
"question": "What specific scene context information should UAV2 share with other UAVs to enhance understanding of the urban area?",
"options": {
"A": "Vehicle movement patterns along parallel roads visible in UAV2",
"B": "Structural details of the circular building visible nearby",
"C": "Spatial arrangement of green spaces across views",
"D": "Traffic density variations captured from multiple UAV views"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The UAV2 image captures the urban layout with roads, moving vehicles, and surrounding buildings, while other UAVs provide additional angles showing housing areas, extended roads, and open green spaces.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628412914992445440-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628412914992445440-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628412914992445440-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628412914992445440-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628412914992445440-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629095239127567616.png",
"question_id": "MDMT_OM_UAV4_19",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the red sedan parked near the curved road in front of small houses in UAV4's view?",
"options": {
"A": "The red sedan now seen driving along the highway adjacent to the forest in UAV2",
"B": "The red sedan parked near a small parking lot beside shops in UAV3",
"C": "The blue pickup truck parked at the curb near the apartment building in UAV1",
"D": "The yellow van passing by the curve in UAV5"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV4 shows a curved residential road surrounded by trees and small houses, with a red sedan parked near the road. UAV2 shows the same red sedan now driving along the highway adjacent to the forest.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629095239127567616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629095239127567616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629095239127567616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629095239127567616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629095239127567616-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623936183309908992.png",
"question_id": "Sim5_why2col_UAV3_2",
"question_type": "Why to Collaborate",
"question": "Why should UAV3 collaborate with other UAVs in this scenario?",
"options": {
"A": "To capture objects in the scene that are currently out of UAV3's field of view",
"B": "To detect lighting variations across different areas in the scene",
"C": "To avoid redundant overlap of information from multiple UAV perspectives",
"D": "To reduce the energy consumption during operation by dividing tasks"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "Current image misses the target due to out-of-view, but other views may capture it.",
"image_description": "UAV3 captures a portion of a suburban area, but key targets are missing in its current view. Collaboration with other UAVs can complement these gaps by accessing views that include the missing objects.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623936183309908992-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623936183309908992-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623936183309908992-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623936183309908992-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623936183309908992-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171352874600448.png",
"question_id": "MDMT_QA_UAV3_15",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV3?",
"options": {
"A": "5",
"B": "1",
"C": "3",
"D": "2"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV3",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171352874600448-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171352874600448-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171352874600448-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171352874600448-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171352874600448-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105298882875904.png",
"question_id": "MDMT_UA_UAV2_5",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV2 usable for target perception tasks?",
"options": {
"A": "No, not usable",
"B": "Yes, usable",
"C": "Yes, highly usable",
"D": "Yes, partially usable"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105298882875904-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105298882875904-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105298882875904-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105298882875904-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105298882875904-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628423268493889280.png",
"question_id": "MDMT_CA_UAV2_40",
"question_type": "Causal Assessment",
"question": "What is the primary cause of reduced visibility in this UAV2 image?",
"options": {
"A": "Partial occlusion by trees near the buildings",
"B": "Motion blur caused by camera instability",
"C": "Strong shadows due to uneven lighting",
"D": "Overlapping objects reducing distinguishability"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows several buildings and roads partially occluded by tall trees, which obstruct the view and reduce visibility. This occlusion is the primary issue affecting perception quality.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628423268493889280-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628423268493889280-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628423268493889280-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628423268493889280-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628423268493889280-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "MDMT_OR_UAV4_31",
"question_type": "Object Recognition",
"question": "Which type of vehicle is most prominently visible near the center of the roadway in this scene?",
"options": {
"A": "A green SUV",
"B": "A red car",
"C": "A blue truck",
"D": "A yellow motorcycle"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image depicts a coastal roadway with multiple vehicles, including cars, SUVs, and trucks. The scene features trees along the water's edge and a parking area near rocky terrain. A red car is prominently visible near the center of the roadway amidst other vehicles.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624108967590171392.png",
"question_id": "MDMT_CA_UAV2_11",
"question_type": "Causal Assessment",
"question": "What is the primary factor affecting accurate object detection in this aerial scene?",
"options": {
"A": "Motion blur caused by high-speed UAV movement",
"B": "Underexposure due to insufficient lighting",
"C": "Partial occlusion caused by tall buildings",
"D": "Distorted perspective from high-altitude camera angle"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows objects such as vehicles and roads partially occluded by tall buildings, impeding object detection in certain areas of the scene.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624108967590171392-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624108967590171392-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624108967590171392-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624108967590171392-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624108967590171392-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629223716469174016.png",
"question_id": "MDMT_UA_UAV4_32",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV4 usable for target perception tasks?",
"options": {
"A": "Yes, usable",
"B": "Yes, highly usable",
"C": "Yes, partially usable",
"D": "Yes, barely usable"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629223716469174016-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629223716469174016-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629223716469174016-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629223716469174016-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629223716469174016-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628426933546515456.png",
"question_id": "Sim5_why2col_UAV2_41",
"question_type": "Why to Collaborate",
"question": "Why should UAV2 collaborate with other UAVs in this environment?",
"options": {
"A": "To overcome occlusion caused by dense tree coverage in the area",
"B": "To ensure accurate mapping of building structures visible from UAV2’s view",
"C": "To compensate for lack of visibility of the road intersections in UAV2’s images",
"D": "To avoid redundancy by reducing the overlap of captured images"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot Fallback)",
"annotation_info": "",
"image_description": "UAV2 captures a landscape with dense trees that occlude key areas like the rock and surrounding terrain. Other UAVs provide complementary views of roads, intersections, and building details less visible from UAV2’s perspective.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628426933546515456-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628426933546515456-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628426933546515456-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628426933546515456-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628426933546515456-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628483886999027712.png",
"question_id": "MDMT_UA_UAV5_42",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV5 usable for target perception tasks?",
"options": {
"A": "Yes, highly usable",
"B": "Yes, partially usable",
"C": "Yes, barely usable",
"D": "No, not usable"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628483886999027712-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628483886999027712-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628483886999027712-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628483886999027712-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628483886999027712-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628230839454745344.png",
"question_id": "Sim5_why2col_UAV3_33",
"question_type": "Why to Collaborate",
"question": "Why should UAV3 collaborate with other UAVs in this scenario?",
"options": {
"A": "To overcome partial occlusion of a historical building seen in UAV3's view",
"B": "To improve mapping accuracy by reducing redundant views",
"C": "To detect objects hidden behind vegetation across UAV perspectives",
"D": "To efficiently identify structural integrity issues in the scene"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "UAV3's view shows partial occlusion of the historical building due to trees; other UAV perspectives provide a clearer view of the structure, enabling complementary information.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628230839454745344-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628230839454745344-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628230839454745344-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628230839454745344-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628230839454745344-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629204982295740416.png",
"question_id": "MDMT_SD_UAV1_28",
"question_type": "Scene Description",
"question": "Which of the following most accurately describes the environmental setting in this scene?",
"options": {
"A": "A commercial area adjacent to a highway and surrounded by grassy terrain",
"B": "A rural landscape with isolated buildings and sparse vehicle traffic",
"C": "An industrial zone with large facilities and heavy vehicle congestion",
"D": "A dense urban area with skyscrapers and crowded pedestrian crossings"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 50.64080056179775, 'y': 11.51685393258427, 'width': 1.1060393258427013, 'height': 1.1235955056179776, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 48.1127106741573, 'y': 14.04494382022472, 'width': 0.9480337078651715, 'height': 0.9831460674157295, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 47.87570224719101, 'y': 15.44943820224719, 'width': 1.3430477528089924, 'height': 1.1235955056179776, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 52.77387640449438, 'y': 12.07865168539326, 'width': 1.3430477528089924, 'height': 1.4044943820224702, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 55.53897471910112, 'y': 10.393258426966293, 'width': 0.9480337078651715, 'height': 1.6853932584269664, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 58.62008426966292, 'y': 11.235955056179774, 'width': 1.264044943820224, 'height': 1.8258426966292163, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 61.3061797752809, 'y': 15.168539325842698, 'width': 0.6320224719101049, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The image depicts a commercial area near a highway with ongoing vehicle traffic. This area includes structures such as a parking lot, a small house surrounded by grassy terrain, and a gas station. There are scattered trees across the terrain, and the location borders a large water body visible in the top part of the image.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629204982295740416-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629204982295740416-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629204982295740416-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629204982295740416-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629204982295740416-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629095239127567616.png",
"question_id": "MDMT_CA_UAV2_19",
"question_type": "Causal Assessment",
"question": "What is the primary cause of detection difficulties for parking lot vehicles in this scene?",
"options": {
"A": "Uneven lighting with shadows across the parking lot",
"B": "Heavy noise from image compression",
"C": "Occlusion by large nearby buildings",
"D": "Severe motion blur due to UAV movement"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a parking lot with uneven lighting conditions caused by shadows from surrounding structures and trees, creating areas of higher difficulty for object detection.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629095239127567616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629095239127567616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629095239127567616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629095239127567616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629095239127567616-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_QA_UAV3_46",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV3?",
"options": {
"A": "1",
"B": "2",
"C": "4",
"D": "3"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV3",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628230839454745344.png",
"question_id": "MDMT_CA_UAV1_33",
"question_type": "Causal Assessment",
"question": "What is the primary cause of reduced visibility in this image taken by UAV1?",
"options": {
"A": "Shadowing effects caused by tall buildings",
"B": "Occlusion from dense tree coverage",
"C": "Low resolution due to high altitude of UAV",
"D": "Overexposure from sunlight on reflective surfaces"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows tall trees in several areas partially occluding parts of buildings and streets, which affects visibility and clarity for perception tasks.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628230839454745344-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628230839454745344-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628230839454745344-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628230839454745344-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628230839454745344-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105298882875904.png",
"question_id": "MDMT_OG_UAV5_5",
"question_type": "Object Grounding",
"question": "Where is the red car positioned relative to the black car in this scene?",
"options": {
"A": "The red car is directly behind the black car in the same lane",
"B": "The red car is ahead of the black car but in a different lane",
"C": "The red car is parked to the right of the black car on the sidewalk",
"D": "The red car is passing the black car in the opposite direction"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The aerial view depicts a road with multiple cars in motion. The red car is driving behind the black car in the same lane toward the top of the frame. Trees, buildings, and green spaces are visible in the surrounding area.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105298882875904-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105298882875904-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105298882875904-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105298882875904-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105298882875904-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628497853920271360.png",
"question_id": "MDMT_SD_UAV1_45",
"question_type": "Scene Description",
"question": "What type of environment is shown in the scene captured by UAV1?",
"options": {
"A": "A dense urban area with skyscrapers and industrial buildings",
"B": "A suburban area featuring mid-rise buildings and organized streets",
"C": "A rural area with fields and scattered homes",
"D": "A recreational park with open green spaces and walking paths"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Poor (2/5) - Blurry, noisy, color and brightness problems.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 79.79283707865167, 'y': 44.10112359550562, 'width': 3.160112359550567, 'height': 2.668539325842694, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 51.1938202247191, 'y': 30.89887640449438, 'width': 1.896067415730343, 'height': 2.106741573033709, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 43.688553370786515, 'y': 14.747191011235955, 'width': 1.1850421348314626, 'height': 1.54494382022472, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 43.530547752808985, 'y': 17.134831460674157, 'width': 1.1060393258427013, 'height': 1.2640449438202275, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "The image shows a suburban area with mid-rise rectangular buildings organized along streets. A circular glass structure stands out among residential spaces, and trees are positioned along the roads and throughout the area.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628497853920271360-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628497853920271360-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628497853920271360-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628497853920271360-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628497853920271360-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624108967590171392.png",
"question_id": "Sim5_what2col_UAV3_11",
"question_type": "What to Collaborate",
"question": "What object-specific information should UAV3 share with other UAVs to improve perception of the traffic scene?",
"options": {
"A": "The positions and speeds of the cars traveling along the central street near the tall building",
"B": "Details of the road signs and markings captured near the intersection roundabout",
"C": "The color and type of cars parked or moving near the residential and suburban area",
"D": "The exact locations of traffic lights visible near the tall glass building and intersections"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV3 provides detailed coverage of multiple cars moving along the central street near a tall building, an area partially blocked or occluded in other UAV views.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624108967590171392-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624108967590171392-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624108967590171392-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624108967590171392-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624108967590171392-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628422473693368320.png",
"question_id": "MDMT_OR_UAV2_39",
"question_type": "Object Recognition",
"question": "Which type of vehicle is prominently visible driving on the central multi-lane road in the scene?",
"options": {
"A": "A yellow sedan",
"B": "A green SUV",
"C": "A white truck",
"D": "A blue motorbike"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a townscape with various buildings and roads. A central multi-lane road has several vehicles visible, including a yellow sedan driving prominently, surrounded by other cars, trees, and urban infrastructure.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628422473693368320-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628422473693368320-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628422473693368320-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628422473693368320-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628422473693368320-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171352874600448.png",
"question_id": "MDMT_UA_UAV2_15",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV2 usable for target perception tasks?",
"options": {
"A": "No, not usable",
"B": "Yes, highly usable",
"C": "Yes, partially usable",
"D": "Yes, barely usable"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171352874600448-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171352874600448-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171352874600448-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171352874600448-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171352874600448-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628484472543340288.png",
"question_id": "MDMT_UA_UAV1_43",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV1 usable for target perception tasks?",
"options": {
"A": "Yes, usable",
"B": "Yes, highly usable",
"C": "Yes, partially usable",
"D": "No, not usable"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV1",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628484472543340288-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628484472543340288-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628484472543340288-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628484472543340288-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628484472543340288-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628483886999027712.png",
"question_id": "Sim5_why2col_UAV2_42",
"question_type": "Why to Collaborate",
"question": "Why is collaboration important for UAV2 based on its current view and annotations?",
"options": {
"A": "To enhance the visibility of distant objects detected in UAV2's image",
"B": "To improve UAV2's ability to monitor moving vehicles efficiently",
"C": "To provide redundancy in case UAV2 fails to capture high-resolution imagery",
"D": "To coordinate the navigation of UAV2 for obstacle avoidance"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "UAV2's image shows distant and small objects, but other UAVs capture larger, clearer perspectives of the same objects.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628483886999027712-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628483886999027712-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628483886999027712-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628483886999027712-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628483886999027712-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628423268493889280.png",
"question_id": "MDMT_CA_UAV5_40",
"question_type": "Causal Assessment",
"question": "What is the main cause of reduced visibility in the lower-right section of this image?",
"options": {
"A": "Overexposure due to direct sunlight",
"B": "Occlusion caused by shadowing from tall buildings",
"C": "Low resolution leading to detail loss",
"D": "Lighting variations creating strong contrasts in the scene"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a scene with visible shadows cast by tall buildings across the area. The lower-right section is affected by occlusion caused by these shadows, reducing the clarity of objects in that part.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628423268493889280-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628423268493889280-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628423268493889280-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628423268493889280-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628423268493889280-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628484472543340288.png",
"question_id": "MDMT_SC_43",
"question_type": "Scene Comparison",
"question": "What difference in object visibility is observed between UAV1 and UAV3?",
"options": {
"A": "UAV1 provides a clearer view of the rock than UAV3 due to its angle",
"B": "UAV3 shows the rock completely hidden by trees while UAV1 shows its full visibility",
"C": "Both UAV1 and UAV3 equally reveal the rock's position without obstructions",
"D": "UAV1 and UAV3 show the rock partially occluded by nearby buildings"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 51.11481741573034, 'y': 24.43820224719101, 'width': 1.6590589887640377, 'height': 1.966292134831459, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 50.56179775280899, 'y': 22.191011235955056, 'width': 1.501053370786508, 'height': 1.9662921348314626, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "UAV1 captures a broad view of the area with the rock fully visible surrounded by open space. UAV3, with its focus closer to the road and parking lot, shows the rock slightly obscured due to the perspective, with less prominence in the scene.",
"missing_models": -12,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628484472543340288-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628484472543340288-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628484472543340288-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628484472543340288-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628484472543340288-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628409917747699200.png",
"question_id": "MDMT_SD_UAV5_35",
"question_type": "Scene Description",
"question": "What is the spatial arrangement of the buildings relative to the main road in this scene?",
"options": {
"A": "Buildings are grouped on the right side of the road, forming a dense cluster",
"B": "Buildings are aligned along both sides of the road with scattered placements",
"C": "Buildings are exclusively on the left side of the road and spaced apart",
"D": "Buildings are located far from the road without direct alignment"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 72.20856741573033, 'y': 26.264044943820224, 'width': 1.7380617977528061, 'height': 2.8089887640449476, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 76.31671348314606, 'y': 30.61797752808989, 'width': 2.370084269662925, 'height': 3.0898876404494366, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 61.93820224719101, 'y': 9.831460674157304, 'width': 1.1850421348314626, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "The scene shows a main road running through a landscape surrounded by grassy fields. Buildings are positioned on both sides of the road and are spaced apart. Trees and other vegetation are primarily clustered on the left side, along with urban elements such as a circular building and residential houses.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628409917747699200-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628409917747699200-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628409917747699200-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628409917747699200-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628409917747699200-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629201000237519360.png",
"question_id": "MDMT_OM_UAV2_27",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the bench near the cluster of houses on the curved street visible in UAV2?",
"options": {
"A": "The bench now visible on the lakeside path in UAV3",
"B": "The bench slightly obscured by trees near the trail entrance in UAV4",
"C": "The bench visible next to the building with the red roof in UAV3",
"D": "The bench now visible near the rocky cliffside in UAV5"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV2 shows a bench near a cluster of houses arranged along a curved street. UAV4 shows the same bench slightly obscured by trees near a trail entrance.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629201000237519360-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629201000237519360-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629201000237519360-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629201000237519360-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629201000237519360-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629093666833511680.png",
"question_id": "MDMT_OR_UAV2_18",
"question_type": "Object Recognition",
"question": "Which type of vehicle is most prominently visible driving in the leftmost lane of the highway?",
"options": {
"A": "A green sedan",
"B": "A black truck",
"C": "A white SUV",
"D": "A red motorcycle"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image depicts a multi-lane highway surrounded by forested terrain and rocky outcroppings. Various vehicles, including sedans, trucks, and SUVs, are visible driving along the road. The leftmost lane prominently shows a green sedan traveling along it.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629093666833511680-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629093666833511680-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629093666833511680-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629093666833511680-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629093666833511680-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628248254973832704.png",
"question_id": "MDMT_CA_UAV4_34",
"question_type": "Causal Assessment",
"question": "What is the primary cause of reduced visibility in this image from UAV4?",
"options": {
"A": "Motion blur caused by fast-moving vehicles",
"B": "Overexposure caused by uneven sunlight",
"C": "Occlusion caused by foreground trees",
"D": "Low resolution due to limited sensor quality"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image from UAV4 shows a residential area with multi-level buildings, roads, and parked cars. Several tall trees are present in the foreground, partially blocking the view of key objects such as the roads and buildings, reducing visibility.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628248254973832704-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628248254973832704-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628248254973832704-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628248254973832704-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628248254973832704-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623936250982352640.png",
"question_id": "MDMT_QA_UAV4_3",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV4?",
"options": {
"A": "3",
"B": "5",
"C": "1",
"D": "2"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623936250982352640-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623936250982352640-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623936250982352640-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623936250982352640-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623936250982352640-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629218507914019328.png",
"question_id": "Sim5_when2col_UAV4_31",
"question_type": "When to Collaborate",
"question": "Should UAV4 collaborate with other UAVs to address need for collaboration due to environmental factors across UAVs?",
"options": {
"A": "Yes, due to poor lighting conditions in multiple views",
"B": "No, lighting is adequate across all UAVs",
"C": "Yes, due to low image resolution in some views",
"D": "No, the environment is clear for all UAVs"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: 1 (Yes)",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629218507914019328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629218507914019328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629218507914019328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629218507914019328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629218507914019328-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629091952438388224.png",
"question_id": "MDMT_OM_UAV1_17",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the cyclist seen riding along the lakeside road in UAV1?",
"options": {
"A": "The cyclist now seen turning onto the side street near the residential area in UAV2",
"B": "The cyclist now visible riding behind a red car on the main road in UAV3",
"C": "The pedestrian standing near the traffic light intersection in UAV4",
"D": "The cyclist now seen stationary near the park entrance in UAV5"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV1 shows a cyclist riding along the lakeside road; UAV2 captures the same cyclist turning onto the side street near the residential area.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629091952438388224-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629091952438388224-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629091952438388224-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629091952438388224-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629091952438388224-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629091952438388224.png",
"question_id": "MDMT_OG_UAV2_16",
"question_type": "Object Grounding",
"question": "What is the spatial relationship between the row of buildings and the gas station in this scene?",
"options": {
"A": "The row of buildings is located directly behind the gas station",
"B": "The row of buildings is opposite the gas station across the main road",
"C": "The row of buildings is adjacent to the gas station on the right side",
"D": "The row of buildings is in the corner surrounded by trees, distant from the gas station"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene shows a gas station positioned near the bottom-right corner of the image, with a row of residential buildings directly behind it. Trees and greenery are present near the residential area, and the main road runs parallel to the river adjacent to the left edge of the scene.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629091952438388224-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629091952438388224-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629091952438388224-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629091952438388224-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629091952438388224-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107791289598720.png",
"question_id": "Sim5_what2col_UAV3_8",
"question_type": "What to Collaborate",
"question": "What scene context information should UAV3 share with other UAVs to enhance the perception of the urban environment?",
"options": {
"A": "The positions and movements of vehicles in residential streets and intersections visible to UAV3",
"B": "Details of the trees and vegetation layout in residential areas, where UAV3 has a clear view",
"C": "Insights on the arrangement of buildings and traffic lights around the roundabout not visible to UAV3",
"D": "The dimensions and alignment of roads that UAV3 can observe but are occluded in the other UAV views"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV3 provides a clear view of residential streets with sparse traffic and intersections, complementing denser traffic observations made by other UAVs in urban and commercial areas.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107791289598720-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107791289598720-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107791289598720-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107791289598720-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107791289598720-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171352874600448.png",
"question_id": "MDMT_UA_UAV3_15",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV3 usable for target perception tasks?",
"options": {
"A": "Yes, highly usable",
"B": "Yes, partially usable",
"C": "No, not usable",
"D": "Yes, barely usable"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV3",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171352874600448-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171352874600448-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171352874600448-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171352874600448-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171352874600448-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624170864769928192.png",
"question_id": "MDMT_OM_UAV5_14",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the pedestrian waiting near the red umbrellas visible in UAV5?",
"options": {
"A": "The pedestrian now walking on the sidewalk near the white building in UAV2",
"B": "The pedestrian now standing near the entrance of the park in UAV3",
"C": "The pedestrian now walking along the side of the fountain in UAV4",
"D": "The pedestrian now standing near a yellow car in UAV1"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV5 shows a pedestrian waiting near red umbrellas in a shaded area; UAV4 reveals the same pedestrian walking alongside the fountain in a much brighter and open environment.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624170864769928192-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624170864769928192-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624170864769928192-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624170864769928192-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624170864769928192-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629223716469174016.png",
"question_id": "MDMT_OM_UAV5_32",
"question_type": "Object Matching",
"question": "In UAV5's view, a black sedan is seen driving uphill along a curved road behind the rocky terrain. Which object in another UAV's view corresponds to this black sedan?",
"options": {
"A": "The black sedan now seen from above, moving downhill along the curved road near UAV3.",
"B": "The stationary black SUV parked near houses in UAV2’s view.",
"C": "The black sedan now positioned at the intersection under the bridge in UAV4’s view.",
"D": "The red vehicle driving toward the parking lot in UAV1’s view."
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV5 shows a black sedan driving uphill along a curved road behind a rocky terrain. UAV3 shows the same black sedan viewed from above as it moves downhill along the same curved road.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629223716469174016-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629223716469174016-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629223716469174016-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629223716469174016-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629223716469174016-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629091952438388224.png",
"question_id": "MDMT_SD_UAV3_17",
"question_type": "Scene Description",
"question": "What is the dominant environmental context of this scene?",
"options": {
"A": "A suburban area with residential houses and open green space",
"B": "A rural area dominated by farmland and sparse development",
"C": "An urban area with skyscrapers and dense traffic",
"D": "A coastal region with beaches and waterfronts"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 0.39501404494382025, 'y': 13.202247191011235, 'width': 1.5800561797752808, 'height': 1.9662921348314626, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 0, 'y': 23.735955056179776, 'width': 1.8170646067415728, 'height': 2.387640449438205, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 21.804775280898873, 'y': 8.707865168539326, 'width': 1.1850421348314626, 'height': 2.3876404494382015, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV4; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The scene depicts a suburban area featuring residential houses with tiled roofs adjacent to a main road. There is open green space with trees and a rocky hill formation. Multiple vehicles are visible on the roads, and the overall environment is well-organized with a mixture of natural and human-made features.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629091952438388224-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629091952438388224-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629091952438388224-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629091952438388224-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629091952438388224-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629125437038686464.png",
"question_id": "Sim5_what2col_UAV1_24",
"question_type": "What to Collaborate",
"question": "What scene context information should UAV1 focus on sharing with other UAVs to improve traffic and terrain understanding?",
"options": {
"A": "Traffic conditions along curved roads that are occluded in other UAV views",
"B": "Details of the elevation changes near the rocky cliff areas",
"C": "Forest density and road visibility near residential areas",
"D": "Road intersection layout connecting the broader highway with local roads"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV1 captures curved residential roads and partial traffic patterns, while other UAVs provide complementary views of cliffs, highways, and neighborhood layouts.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629125437038686464-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629125437038686464-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629125437038686464-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629125437038686464-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629125437038686464-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624114615476825088.png",
"question_id": "MDMT_CA_UAV1_13",
"question_type": "Causal Assessment",
"question": "What is the primary factor affecting object detection quality in this image?",
"options": {
"A": "High noise from sensor malfunction",
"B": "Shadow occlusion cast by nearby trees",
"C": "Low resolution due to UAV altitude",
"D": "Overexposure caused by bright sunlight"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows residential houses, but object detection in some areas is affected by shadows cast by trees in the foreground, causing partial occlusion.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624114615476825088-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624114615476825088-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624114615476825088-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624114615476825088-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624114615476825088-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623937902840830464.png",
"question_id": "MDMT_OM_UAV3_4",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the dark gray sedan traveling in the middle lane in UAV3's view?",
"options": {
"A": "The dark gray sedan now seen from a closer view in UAV2, traveling in the left lane",
"B": "The black SUV now seen from above in UAV1, parked near the intersection",
"C": "The dark gray sedan now seen from the side in UAV4, approaching a group of parked cars",
"D": "The silver sedan now seen from the rear in UAV5, moving through a residential area"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "In UAV3, the dark gray sedan is traveling in the middle lane alongside other vehicles, while UAV4 shows the same car seen from the side as it approaches parked cars. UAV2 provides a closer view of a similar vehicle traveling in the left lane, and UAV5 depicts a silver sedan moving through a residential area.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623937902840830464-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623937902840830464-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623937902840830464-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623937902840830464-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623937902840830464-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1623936183309908992.png",
"question_id": "MDMT_CA_UAV4_2",
"question_type": "Causal Assessment",
"question": "What is the primary cause of reduced visibility in this UAV4 image?",
"options": {
"A": "Overexposure due to bright sunlight causing shadow contrast",
"B": "Motion blur from fast-moving objects within the scene",
"C": "Occlusion by tall objects such as trees and buildings",
"D": "Loss of detail due to low-resolution sensors"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image from UAV4 shows strong shadows cast by sunlight, leading to high contrast between lit and shaded areas. This overexposure impacts visibility of details in shadowed regions.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1623936183309908992-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1623936183309908992-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1623936183309908992-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1623936183309908992-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1623936183309908992-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107811717034496.png",
"question_id": "MDMT_OC_UAV3_9",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV3's perspective?",
"options": {
"A": "14",
"B": "12",
"C": "15",
"D": "10"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107811717034496-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107811717034496-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107811717034496-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107811717034496-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107811717034496-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_OM_UAV1_46",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the cluster of four beige apartment buildings seen near the bottom-right corner of UAV1's view?",
"options": {
"A": "The cluster of four beige apartment buildings now seen near the bottom-left corner of UAV3",
"B": "The four beige apartment buildings seen aligned along the road in UAV4's view",
"C": "The cluster of beige buildings seen near the center of UAV2's view, bordered by palm trees",
"D": "The row of smaller houses overlapping the beige apartment buildings in UAV5's view"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV1 shows a cluster of four beige apartment buildings near the bottom-right corner of the image. In UAV2, the same buildings are visible near the center, bordered by rows of palm trees along the perimeter. UAV3 and UAV4 show other structures in different alignments near the road.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628483886999027712.png",
"question_id": "MDMT_OG_UAV3_40",
"question_type": "Object Grounding",
"question": "What is the spatial position of the large tree nearest to the road curve relative to the white boundary wall?",
"options": {
"A": "The large tree is located inside the boundary wall, near the road bend",
"B": "The large tree is positioned behind and outside the boundary wall",
"C": "The large tree is adjacent to the boundary wall but not near the bend in the road",
"D": "The large tree is situated opposite the boundary wall, across the road"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image depicts a road curving through a landscaped area, bordered by a white boundary wall. Trees are scattered across the area, and one large tree is located close to the curve within the boundary wall.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628483886999027712-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628483886999027712-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628483886999027712-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628483886999027712-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628483886999027712-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628409917747699200.png",
"question_id": "MDMT_SC_35",
"question_type": "Scene Comparison",
"question": "Which UAV perspective provides the clearest view of distinct building shapes and architectural features?",
"options": {
"A": "UAV1 clearly shows the rectangular buildings in a linear arrangement along the road",
"B": "UAV2 provides the clearest view of the white classical building and its landmarks",
"C": "UAV3 best captures the densely arranged residential blocks and their uniform structure",
"D": "UAV4 emphasizes the curved road and sparse arrangement of small houses"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 72.20856741573033, 'y': 26.264044943820224, 'width': 1.7380617977528061, 'height': 2.8089887640449476, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 76.31671348314606, 'y': 30.61797752808989, 'width': 2.370084269662925, 'height': 3.0898876404494366, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 61.93820224719101, 'y': 9.831460674157304, 'width': 1.1850421348314626, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "UAV1 provides a view of rectangular buildings along the main road; UAV2 prominently displays a classical white building and landmarks; UAV3 captures uniformly-arranged residential blocks; UAV4 emphasizes small individual houses near a curved road and trees.",
"missing_models": -12,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628409917747699200-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628409917747699200-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628409917747699200-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628409917747699200-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628409917747699200-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628409917747699200.png",
"question_id": "MDMT_CA_UAV4_35",
"question_type": "Causal Assessment",
"question": "What is the primary cause of reduced object detection accuracy in this image captured by UAV4?",
"options": {
"A": "Occlusion of key objects by trees",
"B": "Blurring due to rapid UAV movement",
"C": "Low resolution caused by sensor limitations",
"D": "Overexposure from direct sunlight"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows key areas obscured by tree foliage which occludes objects of interest, such as vehicles and buildings, impacting detection accuracy.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628409917747699200-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628409917747699200-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628409917747699200-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628409917747699200-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628409917747699200-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105318731299328.png",
"question_id": "Sim5_why2col_UAV3_6",
"question_type": "Why to Collaborate",
"question": "Why is it important for UAV3 to collaborate with other UAVs in this scenario?",
"options": {
"A": "To gain detailed visibility of distant objects that appear smaller in UAV3's view",
"B": "To avoid redundancy by eliminating overlapping fields of view",
"C": "To calculate the speed of vehicles by integrating multiple perspectives",
"D": "To optimize battery usage by reducing UAV flight time"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "UAV3 displays a highway scene with vehicles appearing small and distant, while views from other UAVs show the same vehicles larger and more identifiable.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105318731299328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105318731299328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105318731299328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105318731299328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105318731299328-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628494009272228352.png",
"question_id": "Sim5_who2col_UAV4_44",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV4 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV2",
"B": "None (no need for collaboration)",
"C": "UAV1",
"D": "UAV5"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV4",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV2",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628494009272228352-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628494009272228352-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628494009272228352-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628494009272228352-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628494009272228352-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628410764252718848.png",
"question_id": "MDMT_OC_UAV3_36",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV3's perspective?",
"options": {
"A": "17",
"B": "16",
"C": "14",
"D": "19"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628410764252718848-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628410764252718848-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628410764252718848-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628410764252718848-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628410764252718848-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624111375755501568.png",
"question_id": "MDMT_CA_UAV5_12",
"question_type": "Causal Assessment",
"question": "What is the primary cause of potential object detection issues in this UAV5 image?",
"options": {
"A": "Reflections from bright sunlight on building surfaces",
"B": "Occlusion caused by tall trees near roads",
"C": "Low resolution of the UAV camera sensor",
"D": "Motion blur from rapidly moving vehicles"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The UAV5 image shows a scene with bright sunlight reflecting off the building surfaces. This overexposure might obscure certain details, particularly near reflective areas, posing challenges to object detection.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624111375755501568-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624111375755501568-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624111375755501568-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624111375755501568-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624111375755501568-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624114615476825088.png",
"question_id": "MDMT_QA_UAV4_13",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV4?",
"options": {
"A": "1",
"B": "2",
"C": "3",
"D": "5"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624114615476825088-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624114615476825088-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624114615476825088-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624114615476825088-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624114615476825088-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628422473693368320.png",
"question_id": "MDMT_UA_UAV4_39",
"question_type": "Usability Assessment",
"question": "Is the image captured by UAV4 usable for target perception tasks?",
"options": {
"A": "Yes, usable",
"B": "Yes, partially usable",
"C": "Yes, highly usable",
"D": "Yes, barely usable"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628422473693368320-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628422473693368320-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628422473693368320-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628422473693368320-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628422473693368320-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628422473693368320.png",
"question_id": "Sim5_why2col_UAV1_39",
"question_type": "Why to Collaborate",
"question": "Why should UAV1 collaborate with other UAVs in this specific scene?",
"options": {
"A": "To compensate for object occlusion in UAV1's view using complementary angles",
"B": "To minimize energy consumption across multi-UAV operations",
"C": "To expand the imaging area beyond UAV1's current field of view",
"D": "To improve altitude control in navigation systems"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV1",
"source": "Rule-Based & Model-Based (Few-Shot)",
"annotation_info": "The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "UAV1's view includes occluded buildings and trees, whereas other UAVs provide unobstructed perspectives of the same scene, enabling complementary information extraction.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628422473693368320-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628422473693368320-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628422473693368320-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628422473693368320-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628422473693368320-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629109485061830400.png",
"question_id": "MDMT_OM_UAV3_23",
"question_type": "Object Matching",
"question": "Which object in another UAV’s view corresponds to the yellow vehicle in the left lane of the highway observed in UAV3’s view?",
"options": {
"A": "The yellow vehicle now seen merging onto a curved road in UAV2’s view.",
"B": "The yellow vehicle now stationary behind a row of parked cars in UAV4’s view.",
"C": "The yellow vehicle seen traveling in the middle lane, heading toward an underpass in UAV5’s view.",
"D": "The yellow vehicle now seen parked near the trees on the right side of the road in UAV1’s view."
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV3 shows a yellow vehicle traveling in the left lane of a highway, surrounded by trees on one side and an underpass in the background. UAV2 offers a view of the same yellow vehicle merging onto a curved road, with forest and hills visible nearby.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629109485061830400-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629109485061830400-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629109485061830400-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629109485061830400-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629109485061830400-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628426933546515456.png",
"question_id": "MDMT_QA_UAV5_41",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV5?",
"options": {
"A": "3",
"B": "1",
"C": "5",
"D": "2"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628426933546515456-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628426933546515456-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628426933546515456-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628426933546515456-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628426933546515456-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629223716469174016.png",
"question_id": "MDMT_OC_UAV1_32",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV1's perspective?",
"options": {
"A": "51",
"B": "46",
"C": "48",
"D": "49"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV1",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629223716469174016-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629223716469174016-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629223716469174016-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629223716469174016-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629223716469174016-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629201000237519360.png",
"question_id": "MDMT_SD_UAV2_27",
"question_type": "Scene Description",
"question": "What movement pattern is most likely occurring near the curved residential roads in this scene?",
"options": {
"A": "Vehicles are actively moving along the curves of the roads",
"B": "Vehicles are stationary, parked outside the houses",
"C": "Vehicles are moving in straight lines away from residential roads",
"D": "No vehicles are present near the residential roads"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 56.88202247191011, 'y': 8.98876404494382, 'width': 1.4220505617977466, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 56.40800561797753, 'y': 16.01123595505618, 'width': 0.9480337078651573, 'height': 1.8258426966292127, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 61.22717696629213, 'y': 25, 'width': 1.3430477528089924, 'height': 1.6853932584269664, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 64.46629213483146, 'y': 25, 'width': 2.2120786516853883, 'height': 2.3876404494382015, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 67.38939606741573, 'y': 25.56179775280899, 'width': 1.6590589887640448, 'height': 2.247191011235955, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 55.69698033707865, 'y': 2.6685393258426964, 'width': 1.1060393258427013, 'height': 1.5449438202247197, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 56.961025280898866, 'y': 1.8258426966292134, 'width': 1.027036516853947, 'height': 1.6853932584269666, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 54.59094101123595, 'y': 1.8258426966292134, 'width': 0.9480337078651715, 'height': 1.4044943820224722, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 83.34796348314606, 'y': 0.8426966292134831, 'width': 1.6590589887640448, 'height': 1.5449438202247188, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 86.27106741573033, 'y': 0.8426966292134831, 'width': 1.9750702247191043, 'height': 1.2640449438202248, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 89.82619382022472, 'y': 1.544943820224719, 'width': 1.106039325842687, 'height': 1.1235955056179774, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 94.32935393258427, 'y': 4.073033707865169, 'width': 0.7900280898876417, 'height': 1.5449438202247183, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 91.48525280898876, 'y': 2.247191011235955, 'width': 2.0540730337078656, 'height': 1.685393258426966, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 96.30442415730337, 'y': 3.51123595505618, 'width': 0.8690308988764031, 'height': 0.8426966292134832, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 95.27738764044943, 'y': 3.7921348314606744, 'width': 0.7900280898876417, 'height': 0.9831460674157295, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 96.38342696629213, 'y': 5.337078651685393, 'width': 1.4220505617977466, 'height': 1.6853932584269673, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 98.12148876404494, 'y': 5.617977528089887, 'width': 0.9480337078651644, 'height': 1.2640449438202257, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2, UAV3; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The image depicts a residential area with curved roads lined by houses. Vehicles appear mostly stationary, likely parked outside the houses. Trees are interspersed throughout the area, and a larger roadway runs parallel to the river in the background.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629201000237519360-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629201000237519360-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629201000237519360-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629201000237519360-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629201000237519360-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_QA_UAV1_46",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV1?",
"options": {
"A": "2",
"B": "4",
"C": "3",
"D": "1"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV1",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107791289598720.png",
"question_id": "MDMT_SD_UAV5_8",
"question_type": "Scene Description",
"question": "What is the interaction between the vehicles and the intersection in this scene?",
"options": {
"A": "Vehicles are actively crossing the intersection in various directions",
"B": "Vehicles are stopped at the intersection waiting for their turn",
"C": "Vehicles are parked along the roadside near the intersection",
"D": "Vehicles are absent from the intersection area entirely"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Under shadow, Too small; Perception degradation: Under shadow, Low SNR (noisy), Small target / Long distance; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 0 (No); Collaboration partner: None",
"image_description": "The UAV5 scene presents a multi-lane intersection surrounded by buildings and greenery. Vehicles are seen actively traversing the intersection in multiple directions, with no stationary vehicles visible at the crossing.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107791289598720-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107791289598720-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107791289598720-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107791289598720-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107791289598720-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171377571127296.png",
"question_id": "MDMT_OM_UAV2_16",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the white car seen turning left at the intersection in UAV2?",
"options": {
"A": "The white car now seen traveling straight on the road near the train tracks in UAV3",
"B": "The white car now seen parked near the modern fountain in UAV5",
"C": "The white car now seen driving in front of the row of residential houses in UAV4",
"D": "The white van now seen turning at the intersection near the commercial buildings in UAV2"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV2 shows the white car turning left at a major intersection near the tall black building; UAV3 shows the same white car traveling straight on the road parallel to the train tracks.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171377571127296-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171377571127296-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171377571127296-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171377571127296-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171377571127296-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628422473693368320.png",
"question_id": "Sim5_who2col_UAV5_39",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV5 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV2",
"B": "None (no need for collaboration)",
"C": "UAV1",
"D": "UAV4"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV2",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628422473693368320-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628422473693368320-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628422473693368320-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628422473693368320-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628422473693368320-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624106452866560512.png",
"question_id": "MDMT_OR_UAV4_7",
"question_type": "Object Recognition",
"question": "What type of vehicle is prominently visible near the fountain?",
"options": {
"A": "A blue sedan",
"B": "A red SUV",
"C": "A white truck",
"D": "A yellow taxi"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene depicts a roundabout centered around a fountain. Surrounding the fountain are various vehicles, including a red SUV, a white truck, and other types. Residential houses, trees, and buildings are visible in the background.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624106452866560512-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624106452866560512-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624106452866560512-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624106452866560512-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624106452866560512-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624108954431890688.png",
"question_id": "MDMT_OC_UAV4_10",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV4's perspective?",
"options": {
"A": "34",
"B": "32",
"C": "31",
"D": "33"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624108954431890688-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624108954431890688-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624108954431890688-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624108954431890688-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624108954431890688-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628423268493889280.png",
"question_id": "Sim5_who2col_UAV1_40",
"question_type": "Who to Collaborate",
"question": "Which UAV should UAV1 collaborate with for collaboration partner for complementary perspective in multi-UAV setup?",
"options": {
"A": "UAV2",
"B": "None (no need for collaboration)",
"C": "UAV4",
"D": "UAV3"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV1",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: UAV2",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628423268493889280-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628423268493889280-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628423268493889280-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628423268493889280-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628423268493889280-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628494009272228352.png",
"question_id": "MDMT_SD_UAV4_44",
"question_type": "Scene Description",
"question": "What best describes the spatial arrangement of the buildings relative to the large curving road visible in the scene?",
"options": {
"A": "Buildings are positioned only on one side of the curving road",
"B": "Buildings are scattered randomly far from the curving road",
"C": "Buildings are tightly clustered on both sides of the curving road",
"D": "Buildings form a straight line pattern completely parallel to the curving road"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 69.75948033707866, 'y': 96.62921348314607, 'width': 3.5551264044943736, 'height': 3.3707865168539257, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 98.04248595505618, 'y': 86.65730337078652, 'width': 1.9575140449438209, 'height': 4.213483146067404, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information.",
"image_description": "The scene shows a large curving road surrounded by various buildings on both sides. Dense housing structures cluster near the road, and multiple vehicles are present near this road. Trees and open green spaces are also visible among the urban layout.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628494009272228352-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628494009272228352-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628494009272228352-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628494009272228352-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628494009272228352-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624105318731299328.png",
"question_id": "MDMT_OC_UAV5_6",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV5's perspective?",
"options": {
"A": "13",
"B": "15",
"C": "10",
"D": "14"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV5",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624105318731299328-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624105318731299328-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624105318731299328-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624105318731299328-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624105318731299328-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624108967590171392.png",
"question_id": "MDMT_OP_11",
"question_type": "Observing Posture",
"question": "Which UAV perspective better captures the spatial relationship between fenced properties and adjacent roadways?",
"options": {
"A": "UAV1 provides a clearer view of fenced properties due to its side-angle perspective",
"B": "UAV3 provides the most detailed context of proximity to adjacent roadways",
"C": "UAV2 emphasizes the overall alignment of properties with urban infrastructure and roundabouts",
"D": "UAV4 provides an overhead view that minimizes clarity of individual property features"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 2.291081460674157, 'y': 16.15168539325843, 'width': 1.7380617977528088, 'height': 2.3876404494382046, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 3.160112359550562, 'y': 20.646067415730336, 'width': 1.8960674157303363, 'height': 2.5280898876404514, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 6.004213483146067, 'y': 21.769662921348313, 'width': 3.0811095505617985, 'height': 2.808988764044944, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 0.5530196629213483, 'y': 18.820224719101123, 'width': 1.8170646067415728, 'height': 2.528089887640448, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 0, 'y': 9.831460674157304, 'width': 1.659058988764045, 'height': 2.3876404494382015, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 0, 'y': 14.185393258426968, 'width': 1.5010533707865168, 'height': 1.9662921348314608, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 0, 'y': 16.99438202247191, 'width': 1.4220505617977528, 'height': 1.264044943820224, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 55.80110497237569, 'y': 91.90563886696484, 'width': 3.078137332280974, 'height': 7.997895290712975, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 60.142067876874506, 'y': 93.02815048671403, 'width': 3.4727703235990504, 'height': 6.971849513285974, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 49.96053670086819, 'y': 0.28062790493729717, 'width': 1.3417521704814561, 'height': 1.543453477155135, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV2; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "UAV1 captures an angled view of fenced residential properties and open spaces, focusing on architectural details. UAV2 highlights the urban infrastructure with a broader perspective including a central roundabout. UAV3 shows a detailed view of fenced properties aligned closely with parallel roadways. UAV4 offers a high vertical view focusing more on the general layout rather than specific spatial relationships.",
"missing_models": -12,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624108967590171392-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624108967590171392-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624108967590171392-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624108967590171392-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624108967590171392-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629109485061830400.png",
"question_id": "MDMT_OC_UAV4_23",
"question_type": "Object Counting",
"question": "Based on the image analysis, how many targets (vehicles, pedestrians, bicycles) can be observed in UAV4's perspective?",
"options": {
"A": "27",
"B": "31",
"C": "28",
"D": "29"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Rule-Based from all_samples.json",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629109485061830400-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629109485061830400-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629109485061830400-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629109485061830400-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629109485061830400-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628410764252718848.png",
"question_id": "Sim5_what2col_UAV3_36",
"question_type": "What to Collaborate",
"question": "What specific object information should UAV3 share with other UAVs to improve multi-view perception of the scene?",
"options": {
"A": "Precise positions and movements of the blue car on the main road visible in UAV3's view.",
"B": "Architectural details of the silo structures for better object recognition in other UAVs.",
"C": "Tree locations along the road to align environmental features across all UAV views.",
"D": "Details of vehicle congestion near the circular building visible in UAV2's angle."
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV3 provides a clear view of a blue car moving on the road surrounded by urban structures, while other UAVs offer overlapping views of nearby buildings, roads, and farmland features such as silos and houses.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628410764252718848-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628410764252718848-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628410764252718848-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628410764252718848-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628410764252718848-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628248254973832704.png",
"question_id": "Sim5_what2col_UAV5_34",
"question_type": "What to Collaborate",
"question": "What specific scene context information should UAV5 collaborate on with other UAVs to provide a comprehensive multi-view understanding of the area?",
"options": {
"A": "Details of the green open field area visible in UAV5 to confirm its extent across UAV views",
"B": "Information on building shadows affecting visibility for UAV navigation",
"C": "Traffic density and movement patterns near the roads shown in UAV2 and UAV3",
"D": "The geometrical layout of high-rise buildings across all UAVs"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The first image from UAV5 shows a wide open green field area with shadowed high-rise buildings on one side, while subsequent images display dense urban areas with roads, vehicles, and high-rise buildings providing complementary perspectives on traffic flow and infrastructure.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628248254973832704-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628248254973832704-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628248254973832704-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628248254973832704-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628248254973832704-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628483886999027712.png",
"question_id": "MDMT_SD_UAV1_42",
"question_type": "Scene Description",
"question": "What is the relationship between the parked vehicles and the nearby buildings in this scene?",
"options": {
"A": "Vehicles are parked directly in front of the buildings",
"B": "Vehicles are parked on a road across a grass area from the buildings",
"C": "Vehicles are parked alongside the buildings forming a narrow alleyway",
"D": "Vehicles are parked randomly throughout the scene without a clear relationship to the buildings"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Out of FoV, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance, Out of FoV; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 96.06741573033707, 'y': 11.095505617977528, 'width': 1.501053370786508, 'height': 3.511235955056179, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 98.27949438202246, 'y': 7.303370786516854, 'width': 1.5800561797752835, 'height': 2.808988764044943, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV3; Collaboration reasons: The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "The image shows a multi-lane road with vehicles parked along its edges, separated from nearby residential buildings by a grass area bordered with trees. The buildings are arranged in clusters on one side while the vegetation surrounds the road on the other.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628483886999027712-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628483886999027712-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628483886999027712-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628483886999027712-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628483886999027712-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629200928358985984.png",
"question_id": "Sim5_when2col_UAV5_26",
"question_type": "When to Collaborate",
"question": "Should UAV5 collaborate with other UAVs to address need for collaboration due to environmental factors across UAVs?",
"options": {
"A": "Yes, due to poor lighting conditions in multiple views",
"B": "No, lighting is adequate across all UAVs",
"C": "Yes, due to low image resolution in some views",
"D": "No, the environment is clear for all UAVs"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: 1 (Yes)",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629200928358985984-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629200928358985984-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629200928358985984-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629200928358985984-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629200928358985984-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628494009272228352.png",
"question_id": "Sim5_what2col_UAV5_44",
"question_type": "What to Collaborate",
"question": "What specific object information should UAV5 share with other UAVs to enhance multi-view perception of the urban area?",
"options": {
"A": "Locations and dimensions of residential buildings across UAV views",
"B": "Details of the trees and vegetation along the road from different angles",
"C": "Traffic patterns and vehicle positions at intersections visible from multiple UAVs",
"D": "Positions and shapes of silos near the rural buildings visible only in UAV3's view"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV5 captures an overview of an urban area with roads, vehicles, residential buildings, and intersections. Complementary images show silos, a church, additional greenery, and varied vehicle distributions. Information gaps include vehicular movement at intersections and precise alignments of buildings relative to the roads.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628494009272228352-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628494009272228352-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628494009272228352-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628494009272228352-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628494009272228352-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171377571127296.png",
"question_id": "Sim5_what2col_UAV2_16",
"question_type": "What to Collaborate",
"question": "What traffic-related scene context should UAV2 share with other UAVs to enhance multi-view perception of the intersection area?",
"options": {
"A": "The positions of vehicles across the intersection to predict traffic congestion",
"B": "The layout of road markings and traffic lane divisions visible in UAV2's view",
"C": "The architectural details of nearby buildings to map urban surroundings",
"D": "The pattern of pedestrian movements near the intersection zone"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV2",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV2 highlights the intersection with moving cars and visible lane markings, while other UAVs capture surrounding roads, buildings, and pedestrian crossings. Coordination is required to integrate information about traffic flow and congestion prediction.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171377571127296-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171377571127296-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171377571127296-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171377571127296-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171377571127296-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628426933546515456.png",
"question_id": "MDMT_OG_UAV3_39",
"question_type": "Object Grounding",
"question": "What is the spatial relationship of the large building to the curved road and the dense cluster of trees in this scene?",
"options": {
"A": "The large building is located to the right of the curved road, closer to the dense cluster of trees",
"B": "The large building is positioned to the left of the curved road, separated by sparse trees",
"C": "The large building is directly adjacent to the dense cluster of trees, with no road separating them",
"D": "The large building is situated beyond the curved road, closer to the hillside"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene depicts a large building situated to the left side of a curved road. Sparse trees separate the building from the road on one side, while dense clusters of trees are positioned further ahead near the curve.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628426933546515456-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628426933546515456-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628426933546515456-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628426933546515456-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628426933546515456-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629095239127567616.png",
"question_id": "MDMT_OM_UAV3_19",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the purple sedan driving along the curved road near the apartment building in UAV3?",
"options": {
"A": "The purple sedan now seen stationary in the parking lot in UAV2",
"B": "The purple sedan now seen driving along the highway adjacent to UAV4's view",
"C": "The purple sedan now seen parked beside the small residential area in UAV5",
"D": "The red coupe now seen driving along the same curved road near the apartment building in UAV1"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV3 shows a curved road near an apartment building with a purple sedan driving along it; UAV2 reveals the same purple sedan stationary in the parking lot.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629095239127567616-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629095239127567616-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629095239127567616-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629095239127567616-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629095239127567616-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629096982293769472.png",
"question_id": "Sim5_what2col_UAV1_20",
"question_type": "What to Collaborate",
"question": "What specific scene context information should UAV1 share with other UAVs to improve multi-view perception?",
"options": {
"A": "Traffic density on the main road captured by UAV1 to complement views of intersections from other UAVs.",
"B": "Details of the houses and residential area layout observed from higher altitudes.",
"C": "Spatial information about rocky terrains and their position relative to the roads.",
"D": "Exact locations and movement of vehicles across the intersecting roads near the cliffs."
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV1 provides a clear view of traffic density on a multi-lane highway, while other UAVs capture side roads, intersections, and nearby residential areas. Complementary information includes cliff positions and road layouts.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629096982293769472-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629096982293769472-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629096982293769472-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629096982293769472-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629096982293769472-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624108954431890688.png",
"question_id": "MDMT_OM_UAV4_10",
"question_type": "Object Matching",
"question": "Which object in another UAV's view corresponds to the pedestrian walking along the sidewalk near the tall building in UAV4?",
"options": {
"A": "The pedestrian now seen entering the park near the roundabout in UAV3",
"B": "The pedestrian now standing near the white building across the street in UAV2",
"C": "The pedestrian walking along a fenced property in UAV1’s perspective",
"D": "The pedestrian now crossing the road near the red truck visible in UAV5"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "UAV4 shows a pedestrian walking along the sidewalk near a tall building, while UAV2 presents the same pedestrian standing close to a white building across the street. UAV1, UAV3, and UAV5 depict irrelevant objects like fences, parks, and a crossing pedestrian near a red truck.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624108954431890688-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624108954431890688-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624108954431890688-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624108954431890688-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624108954431890688-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624111375755501568.png",
"question_id": "Sim5_when2col_UAV5_12",
"question_type": "When to Collaborate",
"question": "Should UAV5 collaborate with other UAVs to address need for collaboration due to environmental factors across UAVs?",
"options": {
"A": "Yes, due to poor lighting conditions in multiple views",
"B": "No, lighting is adequate across all UAVs",
"C": "Yes, due to low image resolution in some views",
"D": "No, the environment is clear for all UAVs"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV5",
"source": "Rule-Based (Few-Shot)",
"annotation_info": "Annotation indicates: 1 (Yes)",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624111375755501568-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624111375755501568-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624111375755501568-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624111375755501568-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624111375755501568-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624171377571127296.png",
"question_id": "MDMT_OG_UAV3_15",
"question_type": "Object Grounding",
"question": "Where are the row of identical white buildings with red roofs positioned relative to the curved road leading to the residential houses?",
"options": {
"A": "The white buildings are located directly next to the curved road",
"B": "The white buildings are positioned on the opposite side of the curved road, near the residential houses",
"C": "The white buildings are positioned on the open green space perpendicular to the curved road",
"D": "The white buildings are located between the curved road and the larger rectangular building"
},
"correct_answer": "C",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV3",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The image shows a row of identical white buildings with red roofs located on open green space perpendicular to a curved road that leads to a cluster of residential houses. There are larger buildings visible on the right side of the scene, near a broader road with vehicles.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624171377571127296-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624171377571127296-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624171377571127296-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624171377571127296-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624171377571127296-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628502416425925888.png",
"question_id": "MDMT_OR_UAV4_46",
"question_type": "Object Recognition",
"question": "Which type of object is most prominently visible near the curved road in the foreground?",
"options": {
"A": "A large rock",
"B": "A parked bicycle",
"C": "A pedestrian walking",
"D": "A small tree"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_OU.json",
"uav_id": "UAV4",
"source": "Model-Based (Few-Shot)",
"annotation_info": "",
"image_description": "The scene shows an urban area with multiple buildings, roads, and greenery. In the foreground near a curved road, a large rock is visible surrounded by palm trees. The image captures housing complexes, apartment blocks, streets, and parking areas within a well-maintained neighborhood.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628502416425925888-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628502416425925888-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628502416425925888-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628502416425925888-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628502416425925888-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624107791289598720.png",
"question_id": "Sim5_why2col_UAV1_8",
"question_type": "Why to Collaborate",
"question": "Why is it necessary for UAV1 to collaborate with other UAVs given the scene in the images?",
"options": {
"A": "To ensure visibility of vehicles that may be occluded by buildings or walls",
"B": "To adjust color correction for consistency in all UAV views",
"C": "To capture data from more distant areas not covered by UAV1's position",
"D": "To reduce image overlap between UAV views"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_CD.json",
"uav_id": "UAV1",
"source": "Model-Based (Few-Shot Fallback)",
"annotation_info": "",
"image_description": "The scene in UAV1 shows roads with walls and buildings blocking clear visibility of vehicles and other objects. Subsequent UAV images capture areas occluded in UAV1's view, providing complementary angles.",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624107791289598720-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624107791289598720-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624107791289598720-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624107791289598720-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624107791289598720-UAV5.png"
}
},
{
"sequence_frame": "4-40m-1629091952438388224.png",
"question_id": "MDMT_QA_UAV5_17",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV5?",
"options": {
"A": "5",
"B": "3",
"C": "1",
"D": "2"
},
"correct_answer": "B",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV5",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/4-40m-1629091952438388224-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/4-40m-1629091952438388224-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/4-40m-1629091952438388224-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/4-40m-1629091952438388224-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/4-40m-1629091952438388224-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628410764252718848.png",
"question_id": "MDMT_OP_36",
"question_type": "Observing Posture",
"question": "Which UAV perspective better highlights the spatial relationship between the circular building and the main road?",
"options": {
"A": "UAV1 provides a clearer view due to its higher angle and proximity to the circular building",
"B": "UAV2 provides a clearer view due to its focus on the main road and the surrounding structures",
"C": "Both UAV perspectives equally highlight the spatial relationship",
"D": "Neither UAV perspective adequately shows the relationship between the circular building and the road"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_SU.json",
"uav_id": "Unknown_UAV",
"source": "Model-Based (Few-Shot)",
"annotation_info": "Perception issues: Occluded, Under shadow, Too small; Perception degradation: Object occlusion, Under shadow, Low SNR (noisy), Small target / Long distance; Image quality: Fair (3/5) - Some blur, noticeable noise, slight color issues.; Image usability: Not specified; Collaboration time: 1 (Yes); Collaboration content: [{'x': 7.979283707865169, 'y': 0.1404494382022472, 'width': 0.948033707865167, 'height': 0.7022471910112359, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 10.112359550561797, 'y': 1.8258426966292134, 'width': 1.0270365168539328, 'height': 1.6853932584269666, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}, {'x': 10.033356741573034, 'y': 1.8258426966292134, 'width': 1.1850421348314608, 'height': 1.6853932584269666, 'rotation': 0, 'rectanglelabels': ['More information'], 'original_width': 800, 'original_height': 450}]; Collaboration partner: UAV4; Collaboration reasons: The current image suffers from object occlusion, while the other perspective has a complete view of the object and can provide complementary information., The target appears too small or distant in the current image, but is larger and more visible in the other persepctive.",
"image_description": "UAV1 captures a high-angle view emphasizing the circular building and its direct connection to the main road, while subsequent UAVs show different angles focusing on other buildings and spaces further away from the road.",
"missing_models": -12,
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628410764252718848-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628410764252718848-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628410764252718848-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628410764252718848-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628410764252718848-UAV5.png"
}
},
{
"sequence_frame": "6-40m-1628484472543340288.png",
"question_id": "MDMT_QA_UAV4_43",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV4?",
"options": {
"A": "3",
"B": "1",
"C": "2",
"D": "4"
},
"correct_answer": "A",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV4",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/6-40m-1628484472543340288-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/6-40m-1628484472543340288-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/6-40m-1628484472543340288-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/6-40m-1628484472543340288-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/6-40m-1628484472543340288-UAV5.png"
}
},
{
"sequence_frame": "3-40m-1624111375755501568.png",
"question_id": "MDMT_QA_UAV2_12",
"question_type": "Quality Assessment",
"question": "What is the perception quality assessment score (1-5) for the image captured by UAV2?",
"options": {
"A": "5",
"B": "2",
"C": "1",
"D": "3"
},
"correct_answer": "D",
"dataset_file": "VQA_AeroCollab_PA.json",
"uav_id": "UAV2",
"source": "Rule-Based from JSON",
"annotation_info": "",
"image_description": "",
"uav_paths": {
"UAV1": "Sim_5_UAVs/Samples/UAV1/3-40m-1624111375755501568-UAV1.png",
"UAV2": "Sim_5_UAVs/Samples/UAV2/3-40m-1624111375755501568-UAV2.png",
"UAV3": "Sim_5_UAVs/Samples/UAV3/3-40m-1624111375755501568-UAV3.png",
"UAV4": "Sim_5_UAVs/Samples/UAV4/3-40m-1624111375755501568-UAV4.png",
"UAV5": "Sim_5_UAVs/Samples/UAV5/3-40m-1624111375755501568-UAV5.png"
}
}
]