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 ›› 2008, Vol. 34 ›› Issue (19): 222-224. doi: 10.3969/j.issn.1000-3428.2008.19.075

• 人工智能及识别技术 • 上一篇    下一篇

基于运动检测与运动搜索的多目标跟踪

杨艳芳1,齐美彬2,王 倩3,蒋建国2   

  1. (1. 合肥工业大学应用物理系,合肥 230009;2. 合肥工业大学计算机与信息学院,合肥 230009;3. 安徽省广播电影电视局,合肥230022)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-10-05 发布日期:2008-10-05

Multi-object Tracking Based on Motion Detection and Motion Search

YANG Yan-fang1, QI Mei-bin2, WANG Qian3, JIANG Jian-guo2   

  1. (1. Department of Applied physics, Hefei University of Technology, Hefei 230009; 2. School of Computer and Information, Hefei University of Technology, Hefei 230009; 3. Anhui Provincial Administration of Radio, Film & TV, Hefei 230022)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-10-05 Published:2008-10-05

摘要: 提出一种新的单摄像机多目标跟踪方法,采用全局背景减法得到当前帧所有运动区域,利用kalman滤波器及局部背景减法得到已跟踪目标在当前帧的预测区域,根据全局减法运动区域及预测区域的位置及大小来判断是否有遮挡发生,并用不同匹配方法进行目标跟踪。实验表明,该方法能有效提高单摄像机跟踪对目标合并、遮挡等问题的处理能力。

关键词: 多目标跟踪, 背景减法, kalman滤波器

Abstract: The article puts forward a new method for multi-object tracking with single camera. All of the new object area can be obtained by using the overall background subtraction in current frame, then making use of kalman filter and partial background subtractions, potential object area is obtained .According to the position and size of the new object area and the potential object area, it judges whether there is merger or cover exit, and uses different match method to carry on object tracking. Experimental result shows the algorithms can solve the problems such as merger or cover effectively with single camera, and have strong ability in multi-object tracking.

Key words: multi-object tracking, background subtraction, kalman filter

中图分类号: