基于概率假设密度滤波的多目标跟踪算法  

A Survey of PHD Filter Based on Multi-target Tracking

作　　者：王春平 刘江义 杨文兵 WANG Chunping;LIU Jiangyi;YANG Wenbing(Shijiazhuang Campus of Army Engineering University,Shijiazhuang 050003,China;Army Chongqing Military Department of the 9804 Factory,Qujing 655000,China)

机构地区：[1]陆军工程大学石家庄校区,石家庄050003 [2]陆军重庆军代局驻9804厂军代室,云南曲靖655000

出　　处：《火力与指挥控制》2018年第5期5-10,16,共7页Fire Control & Command Control

摘　　要：概率假设密度(PHD)滤波算法在雷达、红外以及可见光等传感器的多目标跟踪中研究越来越广泛。主要研究了基于PHD滤波的多目标跟踪算法的发展现状,简要介绍了PHD滤波算法的基本原理及其在目标运动模型、多检测情况、杂波分布规律假定等方面的研究进展,概述了PHD滤波的实现方法及研究进展,介绍了PHD滤波的主要应用领域,着眼提高目标跟踪精度和实时性的发展要求提出了PHD滤波的展望。Probability Hypothesis Density （PHD） filter algorithm has been widely used in multipletarget tracking research for radar, infrared and visible light sensors. This paper mainly studies thepresent development situation of mult-target tracking algorithm based on PHD filter. this paper briefly introduces the basic principle of PHD filter and research progress of the algorithm in target dynamic and multiple-detection and clutter distribution assumption. An overview of the implementation methodsof PHD filter and the research progress is given. the main application fields of PHD filter are introduced. Aimed at the improvement the target tracking accuracy and real-time performance,expectation of PHD filter is proposed.

关 键 词：概率假设密度 多目标跟踪 序贯蒙特卡罗 高斯混合 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]