Research on target tracking method based on correlation filter

Aiming at the classical correlation filter target tracking method in the field of computer vision, the content and steps of the correlation filter target tracking method were studied and analyzed, and its shortcomings were summarized. Then, in order to solve the problem of boundary effect in the target tracking method of correlation filtering, the construction method of background aware correlation filter is studied and analyzed. Secondly, the scale estimation method in the process of target tracking is introduced. Finally, around the optimization of the correlation filter target tracking method, the improvement ideas are put forward from four aspects: model update, target redetection, scale estimation and target characteristics.