UAVSwarm Dataset: An Unmanned Aerial Vehicle Swarm Dataset for Multiple Object Tracking

In recent years, with the rapid development of unmanned aerial vehicles (UAV) technology and swarm intelligence technology, hundreds small-scale low-cost UAV constitute swarms carry out complex combat tasks in form ad hoc networks, which brings great threats challenges to low-altitude airspace defense. Security requirements for defense, using visual detection detect track incoming swarms, is premise anti-UAV strategy. Therefore, this study first collected many videos manually annotated a dataset named UAVSwarm tracking; thirteen different scenes more than nineteen types were recorded, including 12,598 images—the number each sequence 3 23. Then, two advanced depth models are used as strong benchmarks, namely Faster R-CNN YOLOX. Finally, state-of-the-art multi-object tracking (MOT) models, GNMOT ByteTrack, conduct comprehensive tests performance verification on evaluation metrics. The experimental results show that has good availability, consistency, universality. can be widely training testing various MOT tasks.