一种改进的粒子滤波跟踪算法的应用研究

The Study of an Improved Tracking Algorithm Based on Particle Filter

粒子滤波跟踪算法计算量大且存在严重的粒子退化现象,因此提出了将均值漂移算法嵌入到粒子滤波的跟踪框架中的算法,该算法可以改善粒子滤波的退化现象,大大减少了算法的运行时间,同时克服了均值漂移算法容易陷入局部最大且无法恢复的缺点。监控系统应用的仿真实验结果表明,该方法具有较强的实时性和鲁棒性。

An approved algorithm was supposed because there have large amount of calculation and serious degeneration phenomenon in the algorithm of particle filter. The algorithm of Mean Shift was embeded into the algorithm of particle filter. The approach can improve the degeneracy of the particle filter and reduce the running time. Meanwhile the new algorithm can overcome shortcoming of Mean Shift which is easy to fall into the local maximum and can not be restored. The result of system simulation show that the new approach supposed has strong real-time and robustness in the monitoring system.