Upload README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# π€ Robot Data Collection Guide
|
| 2 |
+
|
| 3 |
+
A practical guide for collecting clean, diverse, and meaningful datasets to train and evaluate robot models.
|
| 4 |
+
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
## π Why This Matters
|
| 8 |
+
|
| 9 |
+
High-quality data is the foundation for:
|
| 10 |
+
|
| 11 |
+
- Behavior cloning
|
| 12 |
+
- Multi-robot coordination
|
| 13 |
+
- Sim-to-real transfer
|
| 14 |
+
- Sensor fusion and modeling
|
| 15 |
+
|
| 16 |
+
---
|
| 17 |
+
|
| 18 |
+
## π What's Inside
|
| 19 |
+
|
| 20 |
+
- β
Types of data to collect (images, joint states, LiDAR, IMU)
|
| 21 |
+
- β
How to record multi-modal data (ROS 2 bag files)
|
| 22 |
+
- β
Labeling methods (manual, semi-auto)
|
| 23 |
+
- β
Data cleaning and preprocessing
|
| 24 |
+
- β
Storage & formats: `.db3`, `.csv`, `JSON`
|
| 25 |
+
|
| 26 |
+
---
|
| 27 |
+
|
| 28 |
+
## π Tools You Can Use
|
| 29 |
+
|
| 30 |
+
| Tool | Purpose |
|
| 31 |
+
|-----------------|----------------------------------|
|
| 32 |
+
| ROS 2 | Real-time data collection |
|
| 33 |
+
| `rosbag2_py` | Python access to .db3 files |
|
| 34 |
+
| DB Browser | Inspect SQLite `.db3` bags |
|
| 35 |
+
| OpenCV / PIL | Preprocess visual data |
|
| 36 |
+
| Pandas | Work with CSVs |
|
| 37 |
+
| Hugging Face | Store models/datasets online |
|
| 38 |
+
|
| 39 |
+
---
|
| 40 |
+
|
| 41 |
+
## π¦ Example Commands
|
| 42 |
+
|
| 43 |
+
```bash
|
| 44 |
+
# Record robot states + camera stream
|
| 45 |
+
ros2 bag record /joint_states /camera/image_raw /tf
|