UAV-ReID: A Benchmark on Unmanned Aerial Vehicle Re-identification in Video Imagery
Daniel Organisciak, Matt Poyser, Aishah Alsehaim, Shanfeng Hu, Brian K. S. Isaac-Medina, Toby P. Breckon, Hubert P. H. Shum · Proceedings of the 17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications · 2022
As unmanned aerial vehicles (UAV) become more accessible with a growing range of applications, the risk of UAV disruption increases. Recent development in deep learning allows vision-based counter-UAV systems to detect and track UAVs with a single camera. However, the limited eld of view of a single camera necessitates multi-camera congurations to match UAVs across viewpoints { a problem known as re-identication (Re-ID). While there has been extensive research on person and vehicle Re-ID to match objects across time and viewpoints, to the best of our knowledge, UAV Re-ID remains unresearched but challenging due to great dierences in scale and pose. We propose the rst UAV re-identication data set, UAV-reID, to facilitate the development of machine learning solutions in multi-camera environments. UAV-reID has two sub-challenges: Temporally- Near and Big-to-Small to evaluate Re-ID performance across viewpoints and scale respectively. We conduct a benchmark study by extensively evaluating dierent Re-ID deep learning based approaches and their variants, spanning both convolutional and transformer architectures. Under the optimal conguration, such approaches are suciently powerful to learn a well-performing representation for UAV (81.9% mAP for Temporally-Near, 46.5% for the more dicult Big-to- Small challenge), while vision transformers are the most robust to extreme variance of scale.