基于概率假设密度滤波方法的多目标跟踪技术综述

概率假设密度 (Probability hypothesis density, PHD) 滤波方法在多目标跟踪、交通管制、图像处理以及多传感器管理等领域得到了广泛关注. 本文对基于PHD滤波方法的多目标跟踪技术的产生、发展及研究现状进行了综述, 主要包括PHD滤波器、PHD执行方法、峰值提取及航迹提取技术、多传感器多目标跟踪及多传感器管理、 PHD平滑器以及多目标跟踪性能评价指标等, 并对PHD滤波器的相关应用进行介绍. 最后, 基于现有PHD滤波进展, 提出了PHD滤波技术在多目标跟踪领域需要重点关注的若干问题.

A Survey of PHD Filter Based Multi-target Tracking

Probability hypothesis density (PHD) filter has attracted much attention in multi-target tracking, traffic control, image processing, multi-sensor management and other fields. An overview of the emergence, the development and the present research situation of the PHD filter in target tracking is presented here. Special attention is paid to the following areas: PHD filter, its implementation method, the peak and track extraction technology, multi-sensor multi-target tracking, multi-sensor management, PHD smoother, the assessment metrics of multi-target tracking performance, and also the relevant applications. Finally, based on the progress of existing PHD filters, some key issues which need to be focused on for PHD filters in multi-target tracking are introduced.