Research on Video Tracking Algorithm which based on improved CamShift and Particle Kalman Filter

The original CamShift moving target tracking algorithm will appear local maximization when the target is occluded or the background color is close to the target color, resulting in target loss. In order to solve this problem, the CamShift algorithm is improved in this paper, and the particle filter and Kalman filter are fused together. A joint video tracking algorithm (CPKF) based on the fusion of improved CamShift algorithm, Particle Filter and Kalman Filter is proposed. The color feature and texture feature are integrated in CamShift to improve its anti-jamming ability in complex scenes. By improved Camshift algorithm to extract target feature and predict the location of the object, Camshift candidate area was calculated by the Bhattacharyya distance and the similarity of the target area is to determine whether the target obscured, if obscured, using particle filter to track local area, and passing the optimal state estimation to kalman filter to track the global area. If not occluded, Continue tracking with improved Kalman filtering. Experimental results show that the proposed algorithm can effectively track moving targets in complex environments in real time, which is significantly better than the traditional target tracking algorithm.