一种新的基于相似度的目标跟踪算法  

New method for object tracking based on similar and hypothesis testing

作　　者：张杨[1] 郭丽[2] 高立群[1] 

机构地区：[1]东北大学信息科学与工程学院,沈阳110819 [2]天津医学院园林系,天津300203

出　　处：《控制与决策》2011年第12期1900-1903,1908,共5页Control and Decision

基　　金：国家自然科学基会项目(81000639)

摘　　要：研究了视频序列中的目标跟踪问题,提出一种新的基于Chebyshev距离的相似度目标跟踪算法.在跟踪过程中,首先计算参考区域与目标区域间的相似度;然后确定在序列中与被跟踪区域相匹配的区域;最后利用统计检验的方法确定跟踪判断的正确性.该方法可以避免估计概率分布密度函数的不足.运用该算法在运动方式已知和未知的情况下分别与卡尔曼滤波和Camshift算法进行了对比实验,结果表明,所提出算法的跟踪效果有所提高.This paper studies regional tracking in the video sequences. A new method for tracking based on similar measure about Chebyshev distance is presented. The similar measure between the reference object and candidate object can be distinguished in the tracking. Then, the method of hypothesis testing is used for test, and the method of statistical test is used to judge fight tracking or not. Finally, the result of the experimem shows that the method can avoid the probability distribution of the estimated density function, and comparative experiments are carried out to Kalman filtering in moving-known and to Camshift in moving- unknown respectively. The results show that the presented algorithm gets better tracking results.

关 键 词：目标跟踪 相似度 Chebyshev距离 统计检验 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]