基于Mean-shift的粒子滤波算法在遮挡目标跟踪中的应用

Application of Particle Filter Algorithms in Occlusion Object Tracking Based on Mean-shift

针对Mean-shift及粒子滤波对遮挡问题的处理,结合二者的优势做出有效改进。一般情况下利用改进的均值漂移算法跟踪目标,在目标发生严重遮挡情况下,给出一个遮挡因子,采用改进的粒子滤波算法跟踪目标。算法能够对发生遮挡后的目标进行持续、稳健的跟踪,提高算法在遮挡情况下目标跟踪的鲁棒性。

Aiming at the application of mean-shift and particle filter algorithms in the occlusion, an improved algorithm is proposed by combining the advantages. Generally speaking, the improved mean-shift algorithm is used to track objects, when the serious occlusion exists, the occluding factor is given, and the improved particle filter algorithm is adopted to track them. The algorithm can track the targets after the occlusion continually and steadily, meanwhile the robustness is increased.