Tracking algorithm based on spatial progressive matching strategy and optimized correlation

The task of multi-target tracking is to correctly associate the identity of the same target in the two scene scenes. How to improve the accuracy of target tracking is still full of challenges. In this article, we propose a tracking algorithm based on spatial progressive matching strategy and optimized correlation. In it, we divide the targets in the scene according to the area of the target detection frame in the scene. By Prioritizing matching of target groups with a larger target frame area, and then matching target groups with a smaller target frame area is the spatial progressive matching strategy we propose. We noticed that in certain scenes where the target moves too quickly, the traditional intersection-to-union method becomes limited to some extent. Therefore, we substituted it with a circular intersection-to-union ratio method, which is more effective in accurately associating the targets in those scenes.