可提取衍生目标的带标签GM-PHD算法

发布时间：2018-03-12 20:53

  本文选题：概率假设密度滤波　切入点：随机有限集　出处：《光电工程》2016年12期 　论文类型：期刊论文

【摘要】：针对带标签的高斯混合概率假设密度滤波算法无法获取衍生目标的问题,提出一种可以提取衍生目标的带标签GM-PHD算法。首先,通过为高斯项加注标签的方式区别不同的目标,以辨别单个目标及其航迹。其次,在滤波过程中,对每一时刻得到的状态估计值与已形成的航迹标签进行匹配关联,实现航迹维持。最后,通过设置衍生阈值来判断状态估计中是否存在衍生目标以及可能产生的目标个数,为新生目标高斯项和可能的衍生目标高斯项重新分配标签,并创建新的航迹。仿真实验结果表明,与传统的带标签GM-PHD算法相比,在衍生目标存在的情况下,改进算法具有更好的跟踪性能。
[Abstract]:In order to solve the problem of Gao Si's mixed probability assumption that the density filter algorithm can not obtain derived targets, a tagged GM-PHD algorithm is proposed to extract derived targets. Firstly, different targets are distinguished by tagging Gao Si items. In order to distinguish a single target and its track. Secondly, in the process of filtering, the state estimation obtained at each time is matched with the track label that has been formed to achieve track maintenance. The derivative threshold is set to determine whether there are derived targets and the number of possible targets in the state estimation, and to reassign labels to Gao Si item of newborn target and Gao Si term of possible derivative target. The simulation results show that compared with the traditional tagged GM-PHD algorithm, the improved algorithm has better tracking performance than the traditional tagged GM-PHD algorithm.
【作者单位】： 西安工程大学计算机科学学院;德克萨斯州大学奥斯汀分校人类生态系;
【基金】：国家自然科学基金项目(61201118) 陕西省自然科学基础研究计划项目(2016JM6030) 西安工程大学研究生创新基金项目(CX201631);西安工程大学学科建设项目
【分类号】：TN713
，

本文编号：1603258