1 synthetic dataset of a railway forest scene featuring dense per-point semantic labels, synchronized RGB imagery, and ego-motion data. The data is structured for modern perception pipelines, specifically targeting point-cloud semantic segmentation and sensor fusion tasks.
Use Cases
- Train 3D semantic segmentation models using the dense per-point semantic labels
- Develop sensor fusion pipelines by mapping synchronized RGB pixels to LiDAR points
- Test ego-motion estimation algorithms using the provided motion and trajectory data
- Prototype railway obstacle detection systems using the forest scene geometry
Strengths
- Dense per-point semantic labels for 3D point-cloud classification
- Synchronized RGB imagery paired with LiDAR data for sensor fusion
- Ego-motion data included for tracking sensor trajectory
- Synthetic railway forest environment (Scene 1) for infrastructure-specific training