计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (32): 186-188.DOI: 10.3778/j.issn.1002-8331.2008.32.055

• 图形、图像、模式识别 • 上一篇    下一篇

静止背景下基于图割的运动目标检测算法

徐秋平1,2,郭 敏1   

  1. 1.陕西师范大学 计算机科学学院,西安 710062
    2.武警工程学院 教育技术中心,西安 710086
  • 收稿日期:2007-12-10 修回日期:2008-02-27 出版日期:2008-11-11 发布日期:2008-11-11
  • 通讯作者: 徐秋平

Moving object detection algorithm in stationary background based on graph cuts

XU Qiu-ping1,2,GUO Min1   

  1. 1.College of Computer Science,Shaanxi Normal University,Xi’an 710062,China
    2.Instructional Technology Centre of Engineering,College of Armed Police Force,Xi’an 710086,China
  • Received:2007-12-10 Revised:2008-02-27 Online:2008-11-11 Published:2008-11-11
  • Contact: XU Qiu-ping

摘要: 运动目标检测是计算机视觉应用领域中基本而又重要的一步。针对背景差法检测边缘粗糙,存有空洞、噪点的不足,提出一种基于图割的运动目标检测算法。首先把问题转化为能量最小化的组合优化问题,然后构造网络使能量与网络的割的容量相对应,最后利用最大流-最小割算法寻找其最优解。实验结果表明,算法检测精度高、鲁棒性强。

关键词: 运动目标检测, 背景差, 图割, 网络流

Abstract: Moving object detection is an essential and important step in solving issues of computer vision.Aiming at the limitations of traditional background subtraction algorithm that the object edge is coarse,holes and noise exist in the result,a novel method based on graph cuts is proposed.First,the problem is transformed into a combinatorial optimization of energy minimization.Then a network is constructed such that the energies can be related to the capacities of the cuts of the network.Finally,the optimized result is gained by the max-flow min-cuts algorithm.Experimental results show the greater accuracy and stronger robustness of the presented method.

Key words: moving object detection, background subtraction, graph cuts, network flow