Target Tracking Algorithm Based on Kalman Filter and Particle Filter

Abstract

Aiming at the problem that Kalman filter tracking algorithm is no longer accurate in non-Gauss and non-linear case and the Particle filter tracking algorithm costs huge computation,a improved targets tracking algorithm based on Kalman filter and Particle filter was proposed.Firstly,a candidate object was gotten by Kalman tracking algorithm.Then,the tracking result would be verified by Particle filter algorithm when the match threshold is lower than a certain.The improved algorithm used buffer to updated object template to ensure the tracking process continuity,stability and accuracy.Experimental results show that this approach can maintain the efficiency of Kalman algorithm and the powerful ability of Particle filter algorithm,so it is of advanced property.