基于全局时空信息的SMC-PHD多目标跟踪算法  

A Global Space-temporal Information-Based SMC-PHD Multi-target Tracking Algorithm

作　　者：杨峰[1] 王永齐[1] 梁彦[1] 潘泉[1] 

机构地区：[1]西北工业大学自动化学院,陕西西安710072

出　　处：《信息与控制》2014年第3期324-333,共10页Information and Control

基　　金：国家自然科学基金资助项目(61374159;61135001;61374023;61203224);航空科学基金资助项目(20125153027);留金发[2012]3043号

摘　　要：为了实现多目标峰值及航迹联合提取,提出了一种基于全局时空信息的序贯蒙特卡洛概率假设密度(SMC-PHD)多目标峰值提取及航迹提取一体化处理方法.该算法利用粒子空间分布信息将粒子聚拢形成粒子簇,构建历史航迹与粒子簇的匹配关系,同时结合粒子权值信息对粒子标签更新,并依据标签演化特性实现目标峰值及航迹提取.仿真结果表明,所提算法具有稳定的跟踪性能,同时目标信息精度得到显著改善.A global space-temporal information-based sequential Monte Carlo probability hypothesis density （ SMC- PHD） multi-target tracking algorithm is proposed to jointly extract multi-target peaks and tracks. This algorithm assembles particles into multiple particle clusters based on the particles＇ space distribution, constructs relationship between tracks and clusters, updates padicle labels based on particle weights, and extracts multi- target peaks and tracks according to the evolving characteristics of the particles. Simulation results demonstrate that the proposed algorithm provides a stable tracking performance and significantly improves multi-target information extraction accuracy.

关 键 词：多目标跟踪 概率假设密度 峰值提取 航迹管理 全局时空信息 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]