自适应分块多线索融合粒子滤波跟踪方法  

Adaptive fragment multi-cue fusion particle filter tracking method

作　　者：孙晓燕[1,2] 常发亮[1] 

机构地区：[1]山东大学控制科学与工程学院,济南250061 [2]山东建筑大学计算机科学与技术学院,济南250101

出　　处：《控制与决策》2014年第9期1678-1682,共5页Control and Decision

基　　金：国家自然科学基金项目(60975025;61273277);教育部留学回国人员科研启动基金项目(20101174);山东省自然科学基金项目(ZR2011FM032);济南市高校自主创新计划项目(201004002);高等学校博士学科专项科研基金项目(20130131110038)

摘　　要：多线索融合是解决复杂情况下跟踪问题的有效手段,为此提出一种基于自适应分块目标模型的多线索融合粒子滤波跟踪方法.根据目标颜色分布自适应分块建立目标描述模型,可提高对目标初始描述的适应性;采用多线索融合粒子滤波跟踪,在跟踪过程中能根据子块可靠程度动态调整权重,提高对剧烈光照变化、目标姿态变化、遮挡等复杂情况的适应性.实验结果表明,所提出的跟踪方法在多种复杂情况下能准确有效地跟踪目标.Multi-cue fusion is an effective means to solve the tracking problem under many complex conditions. Therefore, a multi-cue fusion particle filter tracking method is proposed, in which the target can be adaptive fragmented. The target model is built based on the color distribution of the target, which can improve the adaptability of the initial description. During the tracking process, the multi-cue fusion particle filter can dynamically adjust the weights of fragments according to the reliability of them, which can improve the flexibility of the tracker to deal with some complex circumstances, such as dramatic light change, target posture change and occluded by other objects. Experimental results show that the proposed method can accurately track the target effectively on many kinds of complex circumstances.

关 键 词：目标跟踪 粒子滤波 自适应分块 多线索融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]