视频检测的颜色直方图目标匹配算法  

Color Histogram Matching Algorithm in Video Detection

作　　者：李坤伦[1] 孙大跃[1] 赵东生[1] 

机构地区：[1]长安大学信息工程学院,西安710061

出　　处：《电脑与信息技术》2008年第5期7-10,共4页Computer and Information Technology

基　　金：基金项目:江苏联网监控(2005074)

摘　　要：文章通过对视频检测系统中运动目标跟踪问题的描述,对颜色直方图、颜色空间和直方图距离进行了介绍。基于RGB颜色模型的直方图,对Euler距离、交集距离、二次方距离以及EMD距离等四种直方图距离进行比较,提出了改进的Euler距离匹配算法,并建立了运动目标的匹配模型。该算法理论上简化了算法复杂度,并去掉了直方图量化的过程。实验中,该算法能满足实时性要求,计算量小,具有很强的鲁棒性。This article through to the video detection system in the movement target tracking question description, and carries on the introduction to the color histogram, color space and the histogram distance. Based on the R.GB color model histogram, to compare the Euler distance, the intersection distance, the quadratic distance and the EMD, I proposed an improvement Euler distance matching algorithm, and have established the movement goal match model. This improvement algorithm simplified algorithin complex, and has removed the histogram quantification process. Experhnents show that the algorithm can satisfy the requirements of real-time, and highly robust.

关 键 词：颜色空间 直方图 目标匹配 直方图距离 Euler距离 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]