Toronto-3D: Mobile LiDAR Point Clouds for Urban Roadway Segmentation

Toronto-3D is a mobile LiDAR point cloud dataset developed by WeikaiTan for the semantic segmentation of urban roadway environments. It provides high-density 3D data captured via mobile mapping systems to support deep learning research in urban scene understanding. The repository was last updated in October 2025.

Use Cases

• Semantic segmentation of urban roadway features using 3D point cloud labels
• Training deep learning models for autonomous vehicle perception in urban environments
• Benchmarking 3D point cloud classification algorithms for mobile mapping data

Strengths

• Mobile LiDAR source providing high-density urban point clouds
• Semantic labels specifically tailored for urban roadway segmentation
• Designed for deep learning point cloud segmentation tasks

Limitations

• Unknown total record count and file size
• Missing information regarding specific label categories and class distribution
• Unknown license terms for commercial or academic use

Provenance

Source

WeikaiTan (GitHub)

Collection Method

Mobile LiDAR sensor data collection and manual/semi-automated annotation

Freshness

Last updated October 2025.

Geography

Toronto, Canada