Target tracking algorithm based on particle filter and mean shift under occlusions

Abstract

A new anti-occlusion method for object tracking is presented to solve the problem that traditional visual tracking algorithms often deviate or lose the targets under occlusion. The motion position of blocked object can be obtained by the further iterative calculation of mean shift algorithm in the particle filter tracking results when the target is occluded, and the approximation and accuracy of tracking results are higher. The particle state of estimation and the mean shift of iteration fused by object state can achieve reliable tracking performance under occlusion and gain the optimal location of object. Experimental results show that the method has strong robustness and error-tolerance to occlusion of tracking objects, and has good performance under complex background.