欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2007, Vol. 24 ›› Issue (4): 501-505.DOI: 10.7523/j.issn.2095-6134.2007.4.015

• 论文 • 上一篇    下一篇

基于核密度相关度量的视频目标跟踪

卢晓鹏 殷学民 邹谋炎   

  1. 中国科学院电子学研究所 北京 100080
    中国科学院研究生院 北京 100039
  • 收稿日期:1900-01-01 修回日期:1900-01-01 发布日期:2007-07-15

An Algorithm of Object Tracking Based on Kernel Density Correlation

LU Xiao-Peng, YIN Xue-Min ZOU Mou-Yan   

  1. Institute of Electronics, Chinese Academy of Sciences, Beijing 100080, China
    Graduate School , Chinese Academy of Sciences, Beijing 100039, China
  • Received:1900-01-01 Revised:1900-01-01 Published:2007-07-15

摘要: 传统的Mean shift 方法采用颜色直方图作为特征,以Bhattacharyya系数作为目标参考模板与当前帧中候选目标间的相似度量,通过迭代寻找距离函数的局部最小值,从而得到当前帧中的目标实际位置。由于颜色直方图仅仅描述了图像中目标的全局颜色分布而忽略了空间位置分布,使得当目标邻域中存在与目标相近似的颜色模式时,算法无法取得理想的跟踪效果。本文提出了基于核密度估计相关的距离度量,在描述参考目标和候选目标时,考虑到诸如颜色、梯度等目标像点的特征区间的同时,融入了目标像点的空间位置信息,使得跟踪算法更加稳健和精确,能够更好适应目标和背景变化复杂的应用场合。

关键词: 目标跟踪, 均值位移, 核密度估计, 核密度相关

Abstract: In traditional Mean shift algorithm ,color histogram is usually used as the features vectors. And the dissimilarity between the referenced targets and the target candidates is expressed by the metric derived from the Bhattacharyya coefficients .The traditional mean shift procedure is used to find the real position of the object through looking for the regional minimum of the distance function iteratively .But there exists some limits because loss of space distribution .To overcome this problem ,the method based on correlation between kernel density estimation of tracking region is proposed. Furthermore, some experiments manifest it can improve accuracy and robustness of this tracking algorithm.

Key words: object tracking, Mean shift, kernel density estimation, kernel density correlation

中图分类号: