光照改变预测的自适应码本模型算法  

Adaptive Codebook model algorithm by illumination change prediction

作　　者：安世全[1] 廖春梅[1] 瞿中[1,2] AN Shi-quan LIAO Chun-mei QU Zhong(College of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China School of Software Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China)

机构地区：[1]重庆邮电大学计算机科学与技术学院,重庆400065 [2]重庆邮电大学软件工程学院,重庆400065

出　　处：《计算机工程与设计》2017年第8期2194-2198,共5页Computer Engineering and Design

基　　金：重庆市高校优秀成果转化基金项目(KJZH14219)

摘　　要：为解决传统的Codebook背景模型算法需设置的参数较多,存在光照变化自适应能力不足以及检测精度不高等问题,利用增加学习率的方法对背景模型进行自适应更新,使其适应不同的光照变化环境;对YUV颜色空间中的Codebook模型进行改进,减少经验参数的设置;在双层Codebook模型基础上,通过增加光照改变预测机制,判断当前像素与码字平均Y分量差值和阈值的大小,使背景模型更加准确。实验结果表明,该算法在光照变化的复杂环境下具有较好的检测率。To resolve these problems of traditional Codebook model algorithm, such that it needs to be set more parameters artifi- cially, it can not adapt to the scene with illumination change and the detection precision is not high, a learning rate was added to update the background model to adapt different lighting environment changes. The model was improved in the YUV color space to reduce the empirical parameters. Based on the two-layer Codebook model, by adding illumination change prediction mechanism, the size of the difference between the current pixels and the codeword＇s average Y component and the threshold was judged, the background model was then more accurate. Experimental results show that the improved algorithm has better detection rate in the complex environment with illumination changes.

关 键 词：Codebook模型 YUV颜色空间 自适应 学习率 光照改变预测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]