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 ›› 2013, Vol. 39 ›› Issue (1): 187-190. doi: 10.3969/j.issn.1000-3428.2013.01.040

• 人工智能及识别技术 • 上一篇    下一篇

一种基于目标分割的实时跟踪方法

陈临强,汪雯斌,何一娜?   

  1. (杭州电子科技大学图形图像研究所,杭州 310018)
  • 收稿日期:2011-10-11 修回日期:2011-12-01 出版日期:2013-01-15 发布日期:2013-01-13
  • 作者简介:陈临强(1962-),男,教授,主研方向:图像处理,计算机视觉;汪雯斌、何一娜,硕士研究生
  • 基金资助:
    浙江省科技计划基金资助项目(2009C03015-4)

A Real-time Tracking Method Based on Object Segmentation

CHEN Lin-qiang, WANG Wen-bin, HE Yi-na   

  1. (Institute of Graphics and Image, Hangzhou Dianzi University, Hangzhou 310018, China)
  • Received:2011-10-11 Revised:2011-12-01 Online:2013-01-15 Published:2013-01-13

摘要: 为减少颜色相似背景在目标跟踪过程中对跟踪结果的影响,提出一种基于目标分割的实时跟踪方法。利用背景差分的方法进行目标分割,使用目标位置、大小及颜色的特征完成目标状态的更新,对连续帧之间的目标实施匹配跟踪,当2个目标重合分离时采用目标大小、颜色特征实现匹配。实验结果表明,该方法匹配跟踪速度较快、跟踪效果较好。

关键词: 遮挡, 多目标跟踪, 主颜色特征, 大小特征, 目标匹配, 前景分割

Abstract: In order to reduce color similar background effect on target tracking process, this paper proposes a real-time target tracking method based on object segmentation. Multiple objects are obtained by using background subtraction method. Each target state can be updated by using the position, size and color characteristic of each target. The similar targets in the consecutive frames can be tracked by using matching targets and when the occluded targets separate, the targets can be tracked well by using this feature. Experimental results show that this method is fast and has a desirable tracking result.

Key words: occlusion, multiple object tracking, dominant color feature, size feature, object matching, foreground segmentation

中图分类号: