基于均值漂移粒子滤波的目标跟踪算法研究

Mean Shift Paticle Filter Based Methods for Target Tracking

针对粒子滤波计算量大的问题,将视觉跟踪领域的均值漂移算法(Mean Shift)与粒子滤波(PF)算法相结合,该算法利用均值漂移算法在重采样之后将粒子收敛到靠近目标真实状态的区域内,改善了传统粒子滤波器的退化现象,减少了算法的运行时间,通过被动跟踪仿真实例,同时使用均值漂移粒子滤波与传统粒子滤波进行跟踪仿真,分析了轨迹跟踪性能,利用均方根误差比较了误差性能。仿真结果表明,Mean Shift PF具有更高的跟踪精度,并且运行时间显著减少。

To reduce the computation burden of particle filter,the Mean Shift and particle filter were combined.The mean shift technique was used to estimates the gradient of the approximated density and moved particles toward the modes of the posterior,leading to a more effective allocation of particles thereupon fewer particles were needed and the computational demand was reduced.The performance were analyzed by comparison with the PF.Simulation results show that MS-PF algorithm can reduce computation burden.Further more, the track accuracy of the algorithm can also be effectively improved.