基于自适应分块的视频目标跟踪方法研究  

Study on vision object tracking based on adaptive object segmentation

作　　者：郝慧娟[1] 许继勇[1] 刘广起[1] 

机构地区：[1]山东省计算中心

出　　处：《计算机应用研究》2010年第9期3579-3580,3590,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60802030);山东省科学院博士基金资助项目(200812)

摘　　要：针对目标跟踪中的遮挡问题,提出了一种基于颜色特征的自适应跟踪算法。该算法利用模糊K均值聚类将目标自适应分块,采用单调递减的核函数对子块进行加权,目标模型的更新只需更新子块的颜色均值,计算量小。在跟踪过程采用目标整体匹配的方法,结合设计的自适应Kalman滤波器,有效地解决了跟踪过程中的目标遮挡问题。实验结果表明,新算法能够实现目标的准确跟踪。In order to solve the object tracking under occlusion,this paper proposed the adaptively tracking algorithm based on color features. Adaptively divided the object using fuzzy K-means clustering algorithm,and weighted with the sub-regions the monotone decreasing kernel function. It updated The object model through the mean value of sub-regions’color,so the calculation was simple. During the object tracking,used the method of integral matching,combining with the adaptive Kalman filter,resolved the object tracking under occlusion effectively. The experiments show that the new algorithm can track the object exactly.

关 键 词：目标跟踪 颜色直方图 自适应分块 KALMAN滤波 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]