README.md

metadata

license: cc-by-4.0
task_categories:
  - object-detection
tags:
  - Autonomous Driving
  - 3D Object Detection
  - Semantic Segmentation
pretty_name: PandaSet
size_categories:
  - 10K<n<100K

PandaSet aims to promote and advance research and development in autonomous driving and machine learning. The first open-source dataset made available for both academic and commercial use, PandaSet combines Hesai’s best-in-class LiDAR sensors with Scale AI’s high-quality data annotation. PandaSet features data collected using a forward-facing LiDAR with image-like resolution (PandarGT) as well as a mechanical spinning LiDAR (Pandar64). The collected data was annotated with a combination of cuboid and segmentation annotation (Scale 3D Sensor Fusion Segmentation).

PandaSet features:

• 48,000 camera images
• 16,000 LiDAR sweeps
• 103 scenes of 8s each
• 28 annotation classes
• 37 semantic segmentation labels
• Full sensor suite: 1x mechanical LiDAR, 1x solid-state LiDAR, 6x cameras, On-board GPS/IMU

Official webpage | Dev-kit

Citation:

@INPROCEEDINGS{9565009,
  author={Xiao, Pengchuan and Shao, Zhenlei and Hao, Steven and Zhang, Zishuo and Chai, Xiaolin and Jiao, Judy and Li, Zesong and Wu, Jian and Sun, Kai and Jiang, Kun and Wang, Yunlong and Yang, Diange},
  booktitle={2021 IEEE International Intelligent Transportation Systems Conference (ITSC)}, 
  title={PandaSet: Advanced Sensor Suite Dataset for Autonomous Driving}, 
  year={2021},
  pages={3095-3101},
  doi={10.1109/ITSC48978.2021.9565009}}

Note: I am not affiliated with the creators of PandaSet (Scale and Hesai).