High-Fidelity 128-Channel LiDAR Dataset For Autonomous Driving

DurLAR is a high-fidelity 128-channel LiDAR dataset. It includes an exemplar subset of 600 frames available for preview, featuring image-based topics such as image_00, image_01, reflectivity, and ambient data. The dataset is intended for autonomous driving and robotics applications.

Use Cases

• Train sensor fusion models using the 128-channel LiDAR point clouds alongside image_00 and image_01 camera data.
• Analyze material properties by correlating LiDAR reflectivity data with ambient light measurements.
• Develop perception algorithms for autonomous vehicles using the high-fidelity 128-channel point cloud data.
• Benchmark 3D object detection and segmentation models on the provided 600-frame exemplar dataset.

Strengths

• Provides a high-fidelity 128-channel LiDAR data source.
• Includes an exemplar subset of 600 frames for direct preview and evaluation.
• Features multiple data modalities including image topics and reflectivity.

Limitations

• The full dataset size, row count, and specific file formats are unknown.
• The geographic and temporal coverage of the data collection is unspecified.
• The license terms for use and redistribution are not provided.

Provenance

Source

Hugging Face dataset uploaded by author 'l1997i'.

Collection Method

Collected using an OS1-128 LiDAR sensor, with intrinsic parameters provided.

Freshness

Last updated on March 20, 2025, with recent additions in March 2025.