Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,130 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# lingbot-depth Dataset
|
| 2 |
+
|
| 3 |
+
Self-curated RGB-D dataset for training [lingbot-depth](https://github.com/lingbot/lingbot-depth), a depth estimation model. Each sample contains an RGB image, raw sensor depth, and ground truth depth.
|
| 4 |
+
|
| 5 |
+
**Total size:** 2.71 TB
|
| 6 |
+
**Depth scale:** millimeters (mm), stored as 16-bit PNG
|
| 7 |
+
**License:** [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/)
|
| 8 |
+
|
| 9 |
---
|
| 10 |
+
|
| 11 |
+
## Sub-datasets
|
| 12 |
+
|
| 13 |
+
| Name | Description |
|
| 14 |
+
|------|-------------|
|
| 15 |
+
| **RobbyReal** | Real-world indoor scenes captured with multiple RGB-D cameras |
|
| 16 |
+
| **RobbyVla** | Real-world data collected during VLA (Vision-Language-Action) robot manipulation tasks |
|
| 17 |
+
| **RobbySim** | Simulated data rendered from two camera viewpoints |
|
| 18 |
+
| **RobbySimVal** | Validation split of simulated data |
|
| 19 |
+
|
| 20 |
+
---
|
| 21 |
+
|
| 22 |
+
## Directory Structure
|
| 23 |
+
|
| 24 |
+
### RobbyReal
|
| 25 |
+
|
| 26 |
+
```
|
| 27 |
+
RobbyReal/
|
| 28 |
+
└── <scene_id>/ # e.g. 00001_01_room
|
| 29 |
+
├── orbbec_335_<seq>/
|
| 30 |
+
│ ├── color/ # RGB images
|
| 31 |
+
│ ├── gtdepth/ # Ground truth depth
|
| 32 |
+
│ ├── rawdepth/ # Raw sensor depth
|
| 33 |
+
│ └── intrinsic.txt
|
| 34 |
+
├── orbbec_335L_<seq>/
|
| 35 |
+
│ ├── color/
|
| 36 |
+
│ ├── gtdepth/
|
| 37 |
+
│ ├── rawdepth/
|
| 38 |
+
│ └── intrinsic.txt
|
| 39 |
+
├── realsense_D415_<seq>/
|
| 40 |
+
│ ├── color/
|
| 41 |
+
│ ├── gtdepth/
|
| 42 |
+
│ ├── rawdepth/
|
| 43 |
+
│ └── intrinsic.txt
|
| 44 |
+
├── realsense_D435_<seq>/
|
| 45 |
+
│ ├── color/
|
| 46 |
+
│ ├── gtdepth/
|
| 47 |
+
│ ├── rawdepth/
|
| 48 |
+
│ └── intrinsic.txt
|
| 49 |
+
└── realsense_D455_<seq>/
|
| 50 |
+
├── color/
|
| 51 |
+
├── gtdepth/
|
| 52 |
+
├── rawdepth/
|
| 53 |
+
└── intrinsic.txt
|
| 54 |
+
```
|
| 55 |
+
|
| 56 |
+
### RobbyVla
|
| 57 |
+
|
| 58 |
+
Data collected during VLA robot manipulation tasks on two robot platforms.
|
| 59 |
+
|
| 60 |
+
```
|
| 61 |
+
RobbyVla/
|
| 62 |
+
├── franka/ # Franka robot arm
|
| 63 |
+
│ └── <seq>/
|
| 64 |
+
│ ├── left_realsense405/
|
| 65 |
+
│ │ ├── color/
|
| 66 |
+
│ │ ├── gtdepth/
|
| 67 |
+
│ │ ├── rawdepth/
|
| 68 |
+
│ │ └── intrinsic.txt
|
| 69 |
+
│ └── right_realsense405/
|
| 70 |
+
│ ├── color/
|
| 71 |
+
│ ├── gtdepth/
|
| 72 |
+
│ ├── rawdepth/
|
| 73 |
+
│ └── intrinsic.txt
|
| 74 |
+
└── ur7e/ # UR7e robot arm
|
| 75 |
+
└── <seq>/
|
| 76 |
+
├── left_realsense405/
|
| 77 |
+
│ ├── color/
|
| 78 |
+
│ ├── gtdepth/
|
| 79 |
+
│ ├── rawdepth/
|
| 80 |
+
│ └── intrinsic.txt
|
| 81 |
+
└── right_realsense405/
|
| 82 |
+
├── color/
|
| 83 |
+
├── gtdepth/
|
| 84 |
+
├── rawdepth/
|
| 85 |
+
└── intrinsic.txt
|
| 86 |
+
```
|
| 87 |
+
|
| 88 |
+
### RobbySim
|
| 89 |
+
|
| 90 |
+
Simulated data rendered from two camera viewpoints: `object_view` and `rrt_view`.
|
| 91 |
+
|
| 92 |
+
```
|
| 93 |
+
RobbySim/
|
| 94 |
+
├── intrinsics.txt
|
| 95 |
+
├── object_view/
|
| 96 |
+
│ └── <scene_id>/<cam_id>/ # e.g. xxxx/00, xxxx/01
|
| 97 |
+
│ ├── 0001_cam0_left.jpg # RGB image
|
| 98 |
+
│ ├── 0001_cam0_depth.png # Ground truth depth
|
| 99 |
+
│ └── 0001_cam0_rmd2c.png # Raw depth
|
| 100 |
+
└── rrt_view/
|
| 101 |
+
└── <scene_id>/<cam_id>/
|
| 102 |
+
├── 0001_cam0_left.jpg
|
| 103 |
+
├── 0001_cam0_depth.png
|
| 104 |
+
└── 0001_cam0_rmd2c.png
|
| 105 |
+
```
|
| 106 |
+
|
| 107 |
+
### RobbySimVal
|
| 108 |
+
|
| 109 |
+
Validation split of simulated data.
|
| 110 |
+
|
| 111 |
+
```
|
| 112 |
+
RobbySimVal/
|
| 113 |
+
└── val_view/
|
| 114 |
+
└── <scene_id>/<cam_id>/ # e.g. xxxx/00, xxxx/01
|
| 115 |
+
├── intrinsics.txt
|
| 116 |
+
├── 0001_cam0_rgb.left.jpg # RGB image
|
| 117 |
+
├── 0001_cam0_depth_left.png # Ground truth depth
|
| 118 |
+
└── 0001_cam0_rawdepth.left.png # Raw depth
|
| 119 |
+
```
|
| 120 |
+
|
| 121 |
+
---
|
| 122 |
+
|
| 123 |
+
## File Description
|
| 124 |
+
|
| 125 |
+
| File/Folder | Description |
|
| 126 |
+
|-------------|-------------|
|
| 127 |
+
| `color/` / `*_rgb*.jpg` | RGB images |
|
| 128 |
+
| `gtdepth/` / `*_depth*.png` | Ground truth depth maps (16-bit PNG, unit: mm) |
|
| 129 |
+
| `rawdepth/` / `*_rawdepth*.png` / `*_rmd2c*.png` | Raw depth from sensor before post-processing (16-bit PNG, unit: mm) |
|
| 130 |
+
| `intrinsic.txt` / `intrinsics.txt` | Camera intrinsic parameters |
|