纹理模型法用于影像纹理分类

Using Texture Model to Classify Image Texture

在分析分形维和马尔可夫随机场两种主要纹理模型法的基础上，提出了先用马尔可夫随机场方法进行纹理分类，再用分形维方法对那些具有相同马尔可夫随机场模型及参数值相近的纹理影像进行分类的方法。该方法不仅能提供出影像纹理的纹理基元，还能给出影像粗糙性、自相似程度，减少了纳伪误差。

Based on the analyses of the character of fractal dimension method and Markov random field(MRF),an image texture classification method is presented.Firstly the MRF method is used,then the fractal dimension method.By this way,we not only can supply the information about texture unit,roughness and self similarity of image texture,but also can reduce the error of recepting false and increase the accuracy of image texture classification.