Vehicle Tracking Based on Particle Filter Using Double Features

Video traffic surveillance is of high interest in the field of intelligent transportation systems and the moving vehicle tracking is an essential technique. Particle filter approximate the optimal Bayesian solution for vehicle tracking as a nonlinear or non-Gaussian system. In this paper a vehicle tracking method based on PF is presented, which combines gray and contour feature particles using fusion algorithm to balance the weights according to the present scene. It is adaptable to the scene because it utilizes the advantage of the proper feature for the present scene. The experiments demonstrate that the proposed method improves the vehicle tracking accuracy and robustness under cluttered scene.