Visual Target Tracking Based on Particle Filter and Local Search

In order to improve the tracking accuracy and robustness,LSPF (Local Search Particle Filter) is proposed that is the visual target tracking algorithm based on the combination of particle filter (PF) and local search (LS). After weight of each sample is obtained by particle filter,local search analysis is applied to each sample to make the samples of small weight converge at that of large weight that overcomes degeneracy problem efficiently.The experimental results show that the average tracking error of standard PF algorithm is 10.89,however,the average tracking error of the new algorithm is only 3.49,and tracking performance is improved evidently.Especially,when the target is interfered,it can be accurately tracked with LSPF algorithm which ensures advantageously the stability of tracking.