Anti-occlusion Video Target Tracking Based on Double Threshold Judgment

In recent years, great progress has been made in moving target tracking methods, but there is still no effective solution to the problem that affects the tracking accuracy, especially the problem that the target is occluded in the tracking process. To solve this problem, this paper proposes an anti-occlusion video target tracking method based on double threshold judgment, which combines Discriminative Scale Space Tracking (DSST) algorithm and Kalman filtering algorithm, according to the change of correlation coefficient P between adjacent templates. Firstly, the target is tracked by DSST algorithm, and the location information and scale information of the target are obtained. In the process of tracking the target, the occlusion threshold \( T_{1} \) is set to determine whether occlusion occurs. If not, the tracking of DSST algorithm is continued. Otherwise, the Kalman tracking algorithm is invoked to track the target during occlusion. At the same time, according to the change of correlation coefficient, the threshold \( T_{2} \) is set to judge whether the occlusion is over or not. If the occlusion is over, the DSST algorithm is returned. Otherwise, Kalman filter algorithm is used to track the occlusion until the end of the tracking. The simulation results show that this method can effectively solve the problem of tracking moving objects when they are occluded.