Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (3): 8-9.

• 博士论坛 • Previous Articles     Next Articles

Object tracking fusion algorithm based on feature color space shift

ZHAO Chun-hui,ZHANG Hong-cai,WANG Yong-zhong   

  1. College of Automation,Northwestern Polytechnical University,Xi’an 710072,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-01-21 Published:2008-01-21
  • Contact: ZHAO Chun-hui

基于特征颜色空间变换的目标跟踪融合算法

赵春晖,张洪才,王永忠   

  1. 西北工业大学 自动化学院,西安 710072
  • 通讯作者: 赵春晖

Abstract: A new object tracking algorithm is presented in this paper.It fused the Kalman filter and mean-shift tracking methods to utilize their merits to track object.Before that,the object feature description should be shifted to LUV from RGB.Then the object feature would be robust in color changing of video.The experiment indicated that the new algorithm is validity and veracity.

Key words: color space shift, object tracking, Kalman filter, mean shift

摘要: 利用LUV色彩空间的特性,提出将RGB色彩空间的目标特征描述转换到LUV色彩空间,从而解决目标表面特征变化造成的目标丢失现象,提高目标跟踪算法的鲁棒性。结合卡尔曼滤波和均值漂移跟踪算法的优点,通过一种判别机制将这两个算法得到的跟踪结果进行融合,提高目标跟踪算法的准确性。通过实验证明了新方法的有效性和准确性。

关键词: 色彩空间转换, 目标跟踪, 卡尔曼滤波, 均值漂移