基于背景减的Mean shift目标跟踪算法  

A Meanshift Object-tracking Algorithm Based on Background Subtraction

作　　者：江绍明[1] 

机构地区：[1]内江师范学院物理与电子信息工程学院,四川内江641100

出　　处：《内江师范学院学报》2010年第12期45-48,共4页Journal of Neijiang Normal University

摘　　要：为了解决初始窗口离跟踪目标较远或受干扰时,容易跟踪失败的问题,提出一种基于背景差的Mean shift跟踪模型的算法.采用背景差提取当前帧运动目标,并在当前帧运动目标位置附近进行Mean shift迭代,以巴氏系数判断当前目标和历史目标的匹配程度,根据匹配结果决定当前帧目标为跟踪目标或新增目标.实验分析,该算法可实现快速、有效目标跟踪.A new Meanshift algorithm for tracking object is put forth based on background subtraction so as to avoid the object-tracking failure arising from cases when the primary window stands too far away from the target being tracked or the target is being disturbed.At first,the moving objects in the current frame are picked up by way of background subtraction;and then Mean shift iteration is done near the location of the moving objects in the current frame;and the Pakistan＇s coefficient was applied to judge the matching degree between the current target and historical target;finally,the results were used to tell whether the target in the current frame is the target being tracked or just a new target.Experimental results show that the new algorithm can help achieve fast and effective object-tracking.

关 键 词：目标跟踪 Mean SHIFT 背景减 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]