均值漂移优化的粒子滤波目标跟踪方法

Particle filter target tracking method optimized by mean shift

为了实现粒子集的有效传播,克服粒子滤波跟踪时的退化问题,提出尺度和方向自适应的均值移动优化粒子滤波目标跟踪算法。用改进的均值移动作为一种优化机制对粒子进行传播,使粒子能够有效分散和聚类,有效解决退化问题。最后将该方法应用到真实图像序列中,实验表明算法在性能和效率上有明显提高。

In order to achieve the effective dissemination of the particle set and overcome the degradation of particle filter tracking, a particle filter target tracking algorithm optimalized by Scale and Orientation Adaptive Mean Shift is proposed. Considering improved mean shift as an optimization mechanism of particle propagation, the particles can be effectively dispersed and clustering, so the degradation problems can be effectively solved. Finally, this method is applied to the real image sequences. Experiments show that the performance and efficiency of the algorithm are significantly better.