基于局部比率二元模式的纹理分类算法  

Texture Classification Based on Local Ration Binary Pattern

作　　者：樊恒[1] 陈仲民[2] 向金海[2] 

机构地区：[1]华中农业大学工学院,武汉430070 [2]华中农业大学信息学院,武汉430070

出　　处：《小型微型计算机系统》2016年第5期1044-1047,共4页Journal of Chinese Computer Systems

基　　金：中央高校基本科研业务费专项资金项目(2662014BQ083;2662014PY052)资助

摘　　要：局部二元模式(Local Binary Pattern,LBP)在纹理分类中具有广泛应用,然而现有的LBP分类方法易受到噪声的影响,鲁棒性并不高.为解决这个问题,提出一种利用像素邻域信息实现纹理图像分类的局部比率二元模式方法.该方法考虑到噪声的随机性及图像中像素的局部相似性,使用像素的局部邻域向量来表示此像素及其强度,能够减小噪声带来的影响,并通过局部比率实现中心点像素与邻域像素的比较.实验表明,提出的方法与经典的LBP方法及几种改进的LBP方法相比,具有较强的分类能力,同时对图像中的噪声具有较好的鲁棒性.Local binary pattern （ LBP） has been widely used in texture classification, however the existing LBP method is easily affected by noise and the robustness is low as well. In this paper, we propose a local ration binary pattern based texture classification method using the neighboring information of pixels. By taking the random of noise and the similarity of pixels in a local region into account, this approach utilizes the local vector to represent the pixel and its intensity, which can reduce the influence of noise. In addition, we achieve comparing the central pixel and its surroundings by the local ration. Experimental results demonstrate that the proposed method has better performance in texture classification compared with classic LBP method and its variants, and meanwhile it is robust to the noise.

关 键 词：局部比率二元模式 局部邻域向量 局部比率 纹理分类 图像处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]