You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

DroneVehicle-C — Corrupted DroneVehicle Test Split

DroneVehicle-C is a derivative of the DroneVehicle dataset (Sun et al., 2022), generated by applying a systematic sensor-degradation corruption pipeline to the original test split. It is intended for benchmarking multimodal RGB+TIR fusion robustness in drone-based vehicle detection, and is released for academic and non-commercial use only in accordance with the CC BY-NC-SA 3.0 license of the original work.

Attribution

Original dataset: DroneVehicle — Sun, Y., Cao, B., Zhu, P., & Hu, Q. "Drone-Based RGB-Infrared Cross-Modality Vehicle Detection via Uncertainty-Aware Learning." IEEE Transactions on Circuits and Systems for Video Technology, 2022.

Corruptions applied by: Saksham Singh Birla, University of Twente (BSc thesis, TSCiT 2025).

Note: An unofficial HuggingFace mirror of DroneVehicle exists at McCheng/DroneVehicle but should not be treated as the authoritative source. The original dataset and its license terms are defined by the VisDrone project at Tianjin University (https://github.com/VisDrone/DroneVehicle).

License

This dataset is released under Creative Commons Attribution-NonCommercial- ShareAlike 3.0 (CC BY-NC-SA 3.0), the same license as the original DroneVehicle dataset distributed by the VisDrone project. Commercial use is prohibited.

Structure

Each .tar.gz contains one corruption condition with rgb/ and ir/ subdirectories. labels.tar.gz contains DOTA-format oriented bounding box annotations shared across all conditions.

Conditions (23 corrupted + 1 clean)

Modality Corruption Severities
RGB gaussian_noise s1, s2, s3
RGB motion_blur s1, s2, s3
RGB brightness_shift s1, s2, s3
RGB low_contrast s1, s2, s3
RGB complete_dropout
TIR sensor_noise s1, s2, s3
TIR blur s1, s2, s3
TIR intensity_shift s1, s2, s3
TIR complete_dropout

Citation

If you use DroneVehicle-C, please cite the original DroneVehicle dataset:

@ARTICLE{sun2022UA_CMDet,
  author  = {Sun, Yiming and Cao, Bing and Zhu, Pengfei and Hu, Qinghua},
  journal = {IEEE Transactions on Circuits and Systems for Video Technology},
  title   = {Drone-Based RGB-Infrared Cross-Modality Vehicle Detection Via
             Uncertainty-Aware Learning},
  year    = {2022},
  volume  = {32},
  number  = {10},
  pages   = {6700--6713},
  doi     = {10.1109/TCSVT.2022.3168279}
}
Downloads last month
139

Paper for Saksham224/DroneVehicle-C