Abstract
A drone monitoring system that integrates deep-learning-based detection andtracking modules is proposed in this work. The biggest challenge in adoptingdeep learning methods for drone detection is the limited amount of trainingdrone images. To address this issue, we develop a model-based droneaugmentation technique that automatically generates drone images with abounding box label on drone's location. To track a small flying drone, weutilize the residual information between consecutive image frames. Finally, wepresent an integrated detection and tracking system that outperforms theperformance of each individual module containing detection or tracking only.The experiments show that, even being trained on synthetic data, the proposedsystem performs well on real world drone images with complex background. TheUSC drone detection and tracking dataset with user labeled bounding boxes isavailable to the public.