Computer Engineering and Applications ›› 2009, Vol. 45 ›› Issue (24): 180-182.DOI: 10.3778/j.issn.1002-8331.2009.24.054

• 图形、图像、模式识别 • Previous Articles     Next Articles

Method for moving objection detection based on mixture Guassian models

LIU Guang-yu,LIU Guo-dong   

  1. Department of Communication and Control Engineering,Jiangnan University,Wuxi,Jiangsu 214122,China
  • Received:2008-05-13 Revised:2008-08-25 Online:2009-08-21 Published:2009-08-21
  • Contact: LIU Guang-yu



  1. 江南大学 通信与控制工程学院,江苏 无锡 214122
  • 通讯作者: 刘光宇

Abstract: This paper studies the detection of moving targets in sequence images.Traditional detection methods can not overcome the problems such as changing background images,illusion mutation and target shadow influence.This paper constructs background model with mixture Guassian distribution.And the illumination information is applied for early processing,then removing shadow with HSV color model.So this method can detect the moving targets fast and precisely.The experimental results verify the real-time and reliability of the method.

摘要: 研究了序列图像中的运动目标检测问题。传统的目标检测方法无法克服背景图像变化、场景光线突变、目标物阴影干扰等问题。采用混合高斯分布对背景进行建模,引入亮度信息进行前序处理,并在此基础上使用颜色模型对残留阴影进行移除,因此能够快速准确地检测出运动目标。通过相关仿真实验,证明了该方法具有可靠性和实时性。

CLC Number: