纹理分析

常用函数：

entropy	Entropy of grayscale image
entropyfilt	Local entropy of grayscale image
rangefilt	Local range of image
stdfilt	Local standard deviation of image

graycomatrix	Create gray-level co-occurrence matrix from image
graycoprops	Properties of gray-level co-occurrence matrix

纹理滤波函数：

在感兴趣像素周围定义一个邻域，计算该邻域的统计特性，并用统计结果作为该感兴趣像素的输出值。

应用举例：

1.检测纹理边缘

I = imread('eight.tif');
imshow(I)

K = rangefilt(I);
figure, imshow(K)

2.统计纹理重复特性

circuitBoard = rot90(rgb2gray(imread('board.tif')));
imshow(circuitBoard);

offsets0 = [zeros(40,1) (1:40)'];
%%Create the GLCMs. Call the graycomatrix function specifying the offsets.
glcms = graycomatrix(circuitBoard,'Offset',offsets0)
%%Derive statistics from the GLCMs using the graycoprops function. The example calculates the contrast and correlation.
stats = graycoprops(glcms,'Contrast Correlation');
%%Plot correlation as a function of offset.
figure, plot([stats.Correlation]);
title('Texture Correlation as a function of offset');
xlabel('Horizontal Offset')
ylabel('Correlation')

plot图在offsets7,15,23和30处出现波峰。观察左图可发现，在垂直方向每隔7个像素有重复的模式出现。
 
 

  
  参考：
 
 
 
 

  
  http://www.mathworks.cn/help/images/texture-analysis-1.html