灰度—梯度共生/环形Gabor滤波纹理分类  

Texture Classification Method Combining Gray Shade-Gradient Intergrowth with Ring Form Gabor Filtering

作　　者：张谢群[1] 

机构地区：[1]黑龙江工程学院科技交流培训中心,黑龙江哈尔滨150050

出　　处：《交通科技与经济》2009年第4期115-117,共3页Technology & Economy in Areas of Communications

摘　　要：解决旋转不变纹理分析问题,提出一种新的旋转不变纹理分类方法:灰度-梯度共生(GGC)/环形Gabor滤波(CGF)联合。将灰度-梯度共生矩阵用于纹理图像的统计分析中。此外,将传统的Gabor滤波器(TGF)改进为环形对称形式,通过CGF信道输出区得旋转不变纹理特征。将两种方法结合起来作为分类纹理依据,提出基于信息增益的特征选择算法。实验结果表明该方法具有较高的分类性能。This paper analyze the revolution invariant texute and puts forward a new classification method combinging gray shade-gradient intergrowth with ring form gabor filtering for it. Statistic analysis of texture image was carried out by adopting shade-gradient intergrowth matrix. In additional, ring symmetry form was obtained by improving traditional Gabor filtering machine（TGF）, revolution invariant texture features was also acquired by the output in CGF information channel subsequently. Combinging two methods metioned aouve to carry out the texutre classification, this paper puts forward character selection computation method based on information increment. The results show the this method has higher classification performance.

关 键 词：纹理分析 纹理分类 纹理图像 GABOR滤波 

分 类 号：P23[天文地球—摄影测量与遥感]