A Method for Tracking Multiple Targets based on Multi-feature Fusion

摘要

Traditional target tracking methods are easy to lose the target in complex background, especially tracking multiple targets. In this paper PSO (Particle Swarm Optimization) algorithm is presented, it is used to calculate the weights of features. and weight the features to get optimal feature, then apply the optimal feature to Meanshift. And, combine the adjustable window algorithm with Meanshift to adapt to the change of the target scale. The experiments' results show that the proposed algorithm has improved the robustness and accuracy than the traditional tracking algorithm in complex background and with multi-target.