论文部分内容阅读
针对当前基于随机集的多扩展目标跟踪算法存在计算量大、量测划分不准确和跟踪误差大的问题,在PHD滤波基础上提出一种基于均值漂移(Mean Shift)迭代的新生未知多扩展目标跟踪算法.首先,对聚类后量测数据进行关联,得到新生目标状态,解决目标新生问题;然后,通过Mean Shift迭代获得目标量测集质心,将扩展目标的多量测问题转化为点量测处理;最后,给出其粒子实现方式.仿真实验表明,所提出的算法可以降低跟踪复杂度,提高跟踪效率,在交叉时刻具有稳定的跟踪性能。
Aiming at the problems of large amount of calculation, inaccurate measurement and large tracking errors, the current multi-extension target tracking algorithm based on random sets has proposed a new unknown multi-extension target based on PHD filter based on mean shift iteration Tracking algorithm.Firstly, we correlate the measured data after clustering to get the new target state and solve the target newborn problem. Then, we obtain the target measurement set centroid by mean shift iteration, Finally, the particle realization method is given.The simulation results show that the proposed algorithm can reduce the tracking complexity, improve the tracking efficiency and have a stable tracking performance at the crossover time.