UpsideDownWash Computer Vision Project

This dataset for object detection features footage of sling-load operations in Norway. Labelled objects include: vehicles, boats, humans, payload, helicopters and various cargo objects. The latest update puts alls samples into the training folder, so you can choose a custom validation split.

The frames are extracted from videos by the YouTube channel: @UpsideDownWash, a big thanks to Marcus for sharing his lessons with others.

Annotations by Marvin Brenner Resolution: 1920x1080, scaled to 960x960 objects: person, cargo, vehicle, boat, helicopter, payload

If you use the dataset for research, please cite: M. Brenner and P. Stütz, "UDW: A Dataset For Automated Monitoring Of Cargo-Handling During Sling-Load Operations," 2023 Seventh IEEE International Conference on Robotic Computing (IRC), Laguna Hills, CA, USA, 2023, pp. 223-230, doi: 10.1109/IRC59093.2023.00044.

@INPROCEEDINGS{10473558, author={Brenner, Marvin and Stütz, Peter}, booktitle={2023 Seventh IEEE International Conference on Robotic Computing (IRC)}, title={UDW: A Dataset For Automated Monitoring Of Cargo-Handling During Sling-Load Operations}, year={2023}, volume={}, number={}, pages={223-230}, keywords={Uncertainty;Helicopters;Transportation;Object detection;Autonomous aerial vehicles;Transformers;Safety;Unmanned aerial vehicle;unmanned helicopter;helicopter;behavior analysis;object detection;dataset;drone}, doi={10.1109/IRC59093.2023.00044}}