USVInland Dataset


USVInland is the first multi-sensor dataset for the unmanned surface vehicle in inland waterways. During a period of 4 mouths, we made use of our self-driving boat to record data from sensors including Lidar, stereo camera, millimeter-wave radar, and INS when sailing both manually and autonomously in inland waterways.

Aiming at the requirements and challenges of perception and localization for USVs in real-world inland waterways, the dataset is developed mainly for three tasks: SLAM/odometry, stereo matching, and water segmentation.

By publishing this dataset, it is hoped that USVInland can support researches on autonomous driving above water surface for better solutions to the challenges caused by the unique sailing scenes, so as to push forward the development of USVs for inland waterways.

Click the items in the sidebar panel to choose the data that you want to download, and apply for downloading on corresponding pages.

If you have any questions or requirements related to our dataset, please email

About Radar

If you are interested in the millimeter wave radar used in the dataset or need to order the radar for research, click to [get more information about the millimeter wave radar] .


如您对此数据集使用的毫米波雷达感兴趣,需购买此数据集所使用的毫米波雷达用于科研事项,可点击 【获取更多毫米波雷达相关信息介绍】


If you want to cite our dataset in your research, you can cite:

  author={Y. {Cheng} and M. {Jiang} and J. {Zhu} and Y. {Liu}},
  journal={IEEE Robotics and Automation Letters},
  title={Are We Ready for Unmanned Surface Vehicles in Inland Waterways? The USVInland Multisensor Dataset and Benchmark},


Interested in USVInland? Watch this video below to see what the data look like!