作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2006, Vol. 32 ›› Issue (24): 175-177. doi: 10.3969/j.issn.1000-3428.2006.24.063

• 人工智能及识别技术 • 上一篇    下一篇

基于改进的均值漂移算法的目标跟踪

马 丽,常发亮,乔谊正,刘增晓   

  1. (山东大学控制科学与工程学院,济南 250061)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2006-12-20 发布日期:2006-12-20

Target Tracking Based on Improved Mean Shift Algorithm

MA Li, CHANG Faliang, QIAO Yizheng, LIU Zengxiao   

  1. (School of Control Science and Engineering, Shandong University, Jinan 250061)
  • Received:1900-01-01 Revised:1900-01-01 Online:2006-12-20 Published:2006-12-20

摘要: 提出了一种基于目标颜色特征的改进的均值漂移算法,对符合颜色模板的目标点不论其在直方图中的概率大小,都赋予相同的最大权值,使目标最大限度地成为密度极值区,以克服干扰影响,并提出了一种分块检测遮挡算法,遮挡期间不更新颜色模板,以保证遮挡后恢复准确的跟踪。实验结果表明该算法具有较强的鲁棒性,能有效实现复杂场景下的目标跟踪。

关键词: 目标跟踪, 均值漂移, 遮挡

Abstract: An improved mean shift algorithm is proposed based on color feature. The target point which accords with the color template has the largest weight regardless of its probability in the histogram, making the target density peak and overcoming the serious clutter and occlusion. Effective occlusion detection method based on sub-block is put forward and the color template isn’t updated to resume the accurate tracking after occlusion. Experimental results indicate it is robust and has good performances under complex background.

Key words: Target tracking, Mean shift, Occlusion