Fusion Image Enhancement for Aerial Target Tracking

In UAV aerial shooting scenes, the quality of image and video data collected in bad scenes is not high, which brings many difficulties to the follow-up UAV tracking accuracy. In this paper, we propose a target tracking algorithm fused with image enhancement algorithms to improve the robustness of UAV tracking in harsh scenarios. Firstly, an adaptive image enhancement module is fabricated for the purpose of boosting the input image’s characteristics. Secondly, the time regularized update strategy is improved to improve the real-time performance of the algorithm. Finally, the effectiveness of our approach is confirmed by testing on a publicly available drone UAV123@10fps dataset.