Spaces:
Sleeping
Sleeping
File size: 9,163 Bytes
adddaea 689227c adddaea 689227c adddaea 689227c adddaea 689227c adddaea 689227c adddaea |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 |
import cv2
import os
import time
import importlib.util
import supervision as sv
from ultralytics import YOLO
config_dir = os.path.abspath(os.path.dirname(__file__))
config_path = os.path.join(config_dir, 'PC_CONFIG.py')
spec = importlib.util.spec_from_file_location("PC_CONFIG", config_path)
PC_CONFIG = importlib.util.module_from_spec(spec)
spec.loader.exec_module(PC_CONFIG)
dir = str(os.path.join(PC_CONFIG.BASE_DIR, "weights", "best_task2.pt"))
class Predictor:
def __init__(self):
# Load a pre-trained yolov8n model
print("dir:",dir)
self.model = YOLO(dir) # replace model here
# self.print_class_ids() # Print class IDs upon initialization
# def print_class_ids(self):
# # Print all class names and their corresponding IDs
# for id, name in enumerate(self.model.names):
# print(f"ID: {id}, Name: {name}")
# def predict_id(self, image_file_path, task_type):
# # Load the image
# image = cv2.imread(image_file_path)
# # Run inference on the image
# results = self.model(image)
# # Print results
# print(results)
# # Show annotation
# self.show_annotation(image, results)
# # Extract class name
# class_name, largest_size, detection_id = None, -1, None
# for result in results: # Assuming 'results' is a list
# print(f"task_type is {task_type}")
# if task_type == "TASK_2":
# for prediction in result.predictions:
# print(prediction)
# class_name = prediction.class_name
# detection_id = prediction.detection_id
# if class_name != "Bullseye":
# break
# else:
# for prediction in result.predictions:
# print(prediction)
# if largest_size == -1 or max(prediction.width, prediction.height) > largest_size:
# largest_size = max(prediction.width, prediction.height)
# class_name = prediction.class_name
# detection_id = prediction.detection_id
# if class_name:
# print("class_name = " + class_name)
# else:
# print("class_name = None")
# return class_name, results, detection_id
def predict_id(self, image_file_path, task_type):
# Load the image
image = cv2.imread(image_file_path)
# Validation for image existence
if image is None:
print(f"Error: Could not read image at {image_file_path}")
return None, None, None
# Check the image size and resize if necessary
if image.shape[0] != 640 or image.shape[1] != 640:
image = cv2.resize(image, (640, 640)) # Resize to 640x640
# Run inference on the image
results = self.model(image) # Directly pass the image
# Print results
print(results)
# Show annotation (using YOLOv8's plotting capabilities)
# results[0].show()
# Extract class name, largest size, and detection ID
class_name, largest_size, detection_id = None, -1, 0
# Check if there are any detections
if results[0].boxes is None or len(results[0].boxes) == 0:
print("No detections found in the image")
return class_name, results, detection_id
boxes = results[0].boxes.xyxy # Get bounding boxes (x1, y1, x2, y2)
scores = results[0].boxes.conf # Get confidence scores
class_ids = results[0].boxes.cls # Get class IDs
# Store all detections with their priority
detections_list = []
for i in range(len(boxes)):
detected_class = results[0].names[int(class_ids[i])]
confidence = float(scores[i])
yolo_class_id = int(class_ids[i])
print(f"Processing detection {i}: {detected_class} (confidence: {confidence:.2f}, class_id: {yolo_class_id})")
if task_type == "TASK_2":
# Check by class name - only set Bullseye to lowest priority
if detected_class.lower() != 'bullseye':
# All non-bullseye detections have equal priority (0), sorted by confidence
print(f" Added to list: {detected_class} with normal priority")
detections_list.append({
'index': i,
'class_name': detected_class,
'confidence': confidence,
'priority': 0 # Equal priority for all non-bullseye detections
})
else:
print(f" Bullseye detected - adding with lowest priority")
detections_list.append({
'index': i,
'class_name': detected_class,
'confidence': confidence,
'priority': -10 # Lowest priority for bullseye
})
else:
# Determine the largest bounding box
box_width = boxes[i][2] - boxes[i][0]
box_height = boxes[i][3] - boxes[i][1]
size = max(box_width, box_height)
detection_id = i
if largest_size == -1 or size > largest_size:
largest_size = size
class_name = detected_class
detection_id = i
# For TASK_2, select detection based on priority, then confidence
if task_type == "TASK_2" and detections_list:
print(f"\nTotal detections found: {len(detections_list)}")
# Sort by priority (descending), then by confidence (descending)
detections_list.sort(key=lambda x: (x['priority'], x['confidence']), reverse=True)
# Print sorted list for debugging
print("Sorted detections:")
for det in detections_list:
print(f" - {det['class_name']}: priority={det['priority']}, confidence={det['confidence']:.2f}")
# Select the highest priority detection
selected = detections_list[0]
class_name = selected['class_name']
detection_id = selected['index']
print(f"\n✓ Selected detection: {class_name} (priority: {selected['priority']}, confidence: {selected['confidence']:.2f})")
if class_name:
print("class_name = " + class_name)
timestamp = int(time.time())
# Save the annotated image
try:
results[detection_id].save(f'../data/annotated_images/{class_name}_{timestamp}.jpg')
except:
print("error in saving photo!")
else:
print("class_name = None")
return class_name, results, detection_id
# def show_annotation(self, image, results):
# # Create supervision annotators
# bounding_box_annotator = sv.BoundingBoxAnnotator()
# label_annotator = sv.LabelAnnotator()
# # Process results from YOLOv8
# detections = []
# for result in results:
# for detection in result.boxes.data: # Accessing YOLOv8's box data
# class_id = int(detection[5]) # Class ID
# x1, y1, x2, y2 = map(int, detection[:4]) # Bounding box coordinates
# score = float(detection[4]) # Confidence score
# # Add to detections
# detections.append({
# "bbox": [x1, y1, x2, y2],
# "confidence": score,
# "class_id": class_id
# })
# # Convert detections to the expected format for supervision
# if detections:
# detections = sv.Detections(
# xyxy=[d["bbox"] for d in detections],
# confidence=[d["confidence"] for d in detections],
# class_id=[d["class_id"] for d in detections]
# )
# # Annotate the image with inference results
# annotated_image = bounding_box_annotator.annotate(scene=image, detections=detections)
# annotated_image = label_annotator.annotate(scene=annotated_image, detections=detections)
# # Display the annotated image
# try:
# cv2.imshow("Annotated Image", annotated_image)
# cv2.waitKey(0) # Wait indefinitely until a key is pressed
# except Exception as e:
# print(f"Error displaying image: {e}")
# finally:
# cv2.destroyAllWindows() # Close all OpenCV windows
# else:
# print("No detections found.")
if __name__ == "__main__":
# Example usage
predictor = Predictor()
# Specify the path to your image
image_file_path = os.path.join(PC_CONFIG.FILE_DIRECTORY, "image-rec", "sample_images", "IMG_9325.jpg")
# Predict and display the class name
predictor.predict_id(image_file_path, "TASK_1")
|