| Video target tracking is a hot research topic in the field of target tracking, which involves many subjects such as image processing, computer vision and pattern recognition. It is widely used in military and civil fields. Video target tracking is to provide the basis for the higher level processing, such as the state of the target's dynamic state, and provide the basis for the subsequent abnormal behavior detection. In this paper, the problem of target tracking algorithm for particle filter is studied, and the corresponding improvement algorithm is proposed in this paper.A particle filter algorithm based on multi-feature fusion is proposed to solve the problem of poor robustness of particle filter algorithm. The algorithm dynamically adjusts the fusion coefficient of each feature to combine the color feature and texture feature adaptively, taking into account the overall information and detail information of the target. Experiments show that the algorithm of particle filter based on adaptive feature fusion can overcome the problem of single feature to the scene of the dependence and the ability of expression is not strong enough. Compared with the traditional particle filter algorithm, the proposed algorithm has higher tracking accuracy and portability.Aiming at the phenomenon of particle degradation in particle filter, an improved particle swarm optimization (PSO) algorithm is introduced, and a particle filtering algorithm (ELPSO-PF) is proposed. In this algorithm, the optimal sampling particle set is optimized by the search of the two stage of the global optimal particle. The sampling particle distribution tends to the high weight region, and the variance of the particle weight is reduced. Experiments show that the particle filter algorithm can effectively overcome the particle degradation phenomenon. Compared with the traditional particle filter algorithm, the proposed algorithm has better tracking accuracy and robustness under the background of clutter. |
PDF Full Text Request