基于光照不变性颜色特征的多层前景检测算法

作者：戴金平,韩建平 日期：2010-03-26/span> 浏览：3932 查看PDF文档

基于光照不变性颜色特征的多层前景检测算法

戴金平,韩建平
（杭州电子科技大学 图形图像研究所，浙江 杭州 310018）

摘要：为了更好地适应复杂背景，提出了一种基于光照不变颜色特征的多层前景检测方法。首先，提出了一种具有光照不变性的颜色模型用以解决因光照引起的颜色变化问题（如阴影抑制）；其次，使用多层背景构造方法建立了动态背景模型；最后，利用有效的权重更新方法解决了背景静止物体运动产生的虚假目标问题。实验结果表明，该前景检测方法能够适应光照变化、背景运动等复杂环境，并能够有效处理虚假目标问题。
关键词：运动目标检测；光照不变性；多层背景构造；虚假目标
中图分类号：TP391文献标识码：A文章编号：1001-4551(2010)02-0028-04

Multi-layer foreground detection based on photometric invariant color
DAI Jin-ping, HAN Jian-ping
(Department of Graphics and Image, Hangzhou Dianzi University, Hangzhou 310018, China)
Abstract: In order to adapt to complex background better, a photometric invariant color based multi-layer foreground detection method was proposed. Firstly, a photometric invariant color model was proposed to solve the color change problem caused by illumination (such as shadow inhibition). Secondly, a dynamic background model was constructed using a layer based method of background modeling; Thirdly, an effective weight update method was used to solve the false object problem caused by the moving of the stationary objects in the background. Experiments results show that, the foreground detection method is able to adapt complex of the environment such as inhibition illuminaition changes, background movement, and it is able to effectively deal with the false object problem.
Key words: moving target detection; photometric invariance; multi-layer background modeling; false target
参考文献(References):
［1］王素玉,沈兰荪.智能视觉监控技术研究进展［J］.中国图象图形学报,2007,12(9)：1505-1514.
［2］万缨,韩毅,卢汉清.运动目标检测算法的探讨［J］.计算机仿真，2006，12(10)：221-225.
［3］杨涛,李静,潘泉,等.一种基于多层背景模型的前景检测算法［J］.中国图象图形学报,2008,13(7):1303-1308.
［4］HEIKKILA J, SILVEN O. A Real-time System for Monitoring of Cyclists and Pedestrians［C］. Proceedings of 2nd IEEE Workshop on Visual Surveillance, Fort Collins, Colorado, USA, 1999.
［5］KIM K, CHALIDABHONGSE T H, HARWOOD D, et al. Real-time foreground-background segmentation using codebook model［J］. Real-Time Imaging,2005,11(3):172-185.
［6］STAUFFER C, GRIMSON W.Learning patterns of activity using real-time tracking［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(8):747-757.
［7］YAO J, ODOBEZ J. Multi-layer Background Subtraction based on Color and Texture[C]//Proc. Int. Conf. Computer Vision and Pattern Recognition,2007:1-8.
［8］贾立好，邹建华.基于新的阈值化方法的背景减法改进［J］.自动化学报,2009,35(4):394-400.
［9］陈祖爵，陈潇君，何鸿.基于改进的混合高斯模型的运动目标检测［J］.中国图象图形学报，2007，12(9):1585-1589.
［10］刘亚，艾海舟，徐光佑.一种基于背景建模的运动目标检测与跟踪算法［J］.信息与控制，2002，31(4):315-319.
［11］李斌,程义民,张玲.基于纹理的运动目标检测［J］.计算机工程与应用,2008,44(7):100-102.