Hybrid tracking based on camshift and template matching
Chunbo Xiu, Ruosi Wang · 2017
In the process of the target tracking, the real-time target tracking can not be realized due to the Interference of the other factors, leading to tracking failure. A hybrid tracking algorithm combining camshift algorithm and template matching algorithm is proposed in this paper. The target tracking is achieved by camshift algorithm. The template matching algorithm is used to find the target again if the target is lost. The specific algorithm is that the tracking target which is used as the template image is selected in the first frame. When each frame in a video stream is traced by camshift algorithm, the Bhattacharyya coefficient is used to judge the tracking result. If the target is missing, the target can be re-found using template matching and be continued to track using camshift algorithm. The experimental results show that the shortcomings of camshift algorithm can be made up by the algorithm in this paper. The tracking effect is more obvious and the tracking integrity is more perfect can be achieved by the hybrid tracking.