一种基于直方图和冗余匹配的实时目标跟踪算法  

A Robust Real-time Moving Object Tracking Algorithm Based on Histogram and Redundancy Match

作　　者：胡明昊[1] 杨文杰[1] 任明武[1] 杨静宇[1] 

机构地区：[1]南京理工大学计算机系,南京210094

出　　处：《计算机工程》2006年第12期202-204,共3页Computer Engineering

摘　　要：基于视觉的目标跟踪技术难点在于完成帧与帧之间的快速且可靠的目标匹配,以目标的直方图作为特征,具有很好的鲁棒性。但是直方图无法反映目标的空间信息,会导致错误定位于伪目标,且无法精确定位目标。针对以上问题,文章提出了一种稳健的运动目标跟踪算法,在以直方图作为目标特征的同时,结合冗余模板信息,达到验证候选目标的可靠性和辅助目标精确定位的目的。实地测试结果验证了该算法的有效性。The difficulty in visual tracking is how to match targets from frame to frame quickly and reliably, Histogram character has good stability in tracking. But it can＇t describe the any space information of object, which leads to find pseudo-object. Then, it can＇t precisely position the object only with the comparability of histogram. In order to resolve the above problems, the paper proposes a robust moving object tracking algorithm. It uses the histogram as object feature and combines the template information for validating the reliability of candidate objects and providing helpful information for positioning the object precisely. The experiment results show that the algorithm is practical and efficient.

关 键 词：目标跟踪 均值偏移 直方图 模板匹配 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]