The ONCE dataset is a large-scale autonomous driving dataset with 2D&3D object annotations.

• 1 Million LiDAR frames, 7 Million camera images

• 200 km² driving regions, 144 driving hours

• 15k fully annotated scenes with 5 classes (Car, Bus, Truck, Pedestrian, Cyclist)

Diverse environments (day/night, sunny/rainy, urban/suburban areas)


The data acquisition vehicle is equipped with one 40-beam LiDAR on the top and seven cameras.



2021/05/18 Initial release.

2021/06/30 Dataset toolkit release.

2021/07/01 The challenge website has been released at CodaLab.

2021/07/14 Data download links have been updated.

2021/08/05 Calibration data updated, all info files should re-download.

Our paper is accepted to NeurIPS 2021 datasets track

We successfully hold the 3D object detection challenge for ICCV 2021 SSLAD Workshop

We open a permanent test leaderboard at codalab

2022/03/28 New annotation on 3D lanes with ONCE dataset are available on ONCE-3Dlanes

2023/02/28 New test leaderboard update due to codalab site update codalab


If you find this dataset useful, please cite the following paper:

title={One Million Scenes for Autonomous Driving: ONCE Dataset},
author={Mao, Jiageng and Niu, Minzhe and Jiang, Chenhan and Liang, Xiaodan and Li, Yamin and Ye, Chaoqiang and Zhang, Wei and Li, Zhenguo and Yu, Jie and Xu, Chunjing and others}, 
journal={arXiv preprint arXiv:2106.11037}, 


Huawei Corporation.