一种改进的高斯混合背景模型算法及仿真  

Simulation of an Improved Gaussian Mixture Model for Background Subtraction

作　　者：陈振华[1] 周锐锐[1] 李光伟[2] 毕笃彦[1] 

机构地区：[1]空军工程大学工程学院航空电子工程系,陕西西安710038 [2]空军装备研究院雷达所,北京100085

出　　处：《计算机仿真》2007年第11期190-193,共4页Computer Simulation

摘　　要：高斯混合模型广泛应用于基于背景建模的运动目标检测中,高斯混合模型参数估计和更新算法影响到背景模型的性能。文中对传统的高斯混合背景模型进行了改进,针对背景局部运动、活动阴影等问题,采用混合色彩值抑制阴影,在背景更新中引入一个"前景支撑映射"(Foreground Support Map,FSB),较好地解决了背景模型的提取、更新、背景扰动、外界光照变化等问题。实验结果证明,实验结果验证了该方法的有效性和在复杂背景变化下的鲁棒性。Gaussians mixture model （GMM） has been widely used for robustly modeling complicated backgrounds. The performance of GMM can be greatly improved by tackling several practical issues. This paper develops a modificative algorithm using Gaussian mixture probability density. A mixed color space is used to detect and suppress moving cast shadow. A set of counter called the ＂Foreground Support Map＂ is exploited. The modificative algorithm effectively deals with the problems of the abstraction and update of background, background disturbance, illumination change etc. Experimental results show that the presented method can achieve much better results - even outperforming other state - of - the - art methods.

关 键 词：背景模型 混合高斯模型 背景更新 目标检测 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]