基于灰度共生矩阵的图象纹理分析
赞:http://wenku.baidu.com/link?url=XEPw0bkbEjIbsREQEiYFuBxzb2w2Za017GHL8TD6f_Gpe5WKo1PHzV2tXdaHKbrEDEaYt3W3CjFoJ0uGsLMgsvFhU_mjtQxF57CfdgsSoTy
http://www.cnblogs.com/rong86/p/3695621.html
%**************************************************************************
% 图像检索——纹理特征
%基于共生矩阵纹理特征提取,d=1,θ=0°,45°,90°,135°共四个矩阵
%所用图像灰度级均为256
%参考《基于颜色空间和纹理特征的图像检索》
%function : T=Texture(Image)
%Image : 输入图像数据
%T : 返回八维纹理特征行向量
%**************************************************************************
% function T = Texture(Image)
Gray = imread('d:\result5.bmp');
[M,N,O] = size(Gray);
M = 128;
N = 128;
%--------------------------------------------------------------------------
%1.将各颜色分量转化为灰度
%--------------------------------------------------------------------------
% Gray = double(0.3*Image(:,:,1)+0.59*Image(:,:,2)+0.11*Image(:,:,3));
%--------------------------------------------------------------------------
%2.为了减少计算量,对原始图像灰度级压缩,将Gray量化成16级
%--------------------------------------------------------------------------
for i = 1:M
for j = 1:N
for n = 1:256/16
if (n-1)*16<=Gray(i,j)&Gray(i,j)<=(n-1)*16+15
Gray(i,j) = n-1;
end
end
end
end
%--------------------------------------------------------------------------
%3.计算四个共生矩阵P,取距离为1,角度分别为0,45,90,135
%--------------------------------------------------------------------------
P = zeros(16,16,4);
for m = 1:16
for n = 1:16
for i = 1:M
for j = 1:N
if j<N&Gray(i,j)==m-1&Gray(i,j+1)==n-1
P(m,n,1) = P(m,n,1)+1;
P(n,m,1) = P(m,n,1);
end
if i>1&j<N&Gray(i,j)==m-1&Gray(i-1,j+1)==n-1
P(m,n,2) = P(m,n,2)+1;
P(n,m,2) = P(m,n,2);
end
if i<M&Gray(i,j)==m-1&Gray(i+1,j)==n-1
P(m,n,3) = P(m,n,3)+1;
P(n,m,3) = P(m,n,3);
end
if i<M&j<N&Gray(i,j)==m-1&Gray(i+1,j+1)==n-1
P(m,n,4) = P(m,n,4)+1;
P(n,m,4) = P(m,n,4);
end
end
end
if m==n
P(m,n,:) = P(m,n,:)*2;
end
end
end
%%---------------------------------------------------------
% 对共生矩阵归一化
%%---------------------------------------------------------
for n = 1:4
P(:,:,n) = P(:,:,n)/sum(sum(P(:,:,n)));
end
%--------------------------------------------------------------------------
%4.对共生矩阵计算能量、熵、惯性矩、相关4个纹理参数
%--------------------------------------------------------------------------
H = zeros(1,4);
I = H;
Ux = H; Uy = H;
deltaX= H; deltaY = H;
C =H;
for n = 1:4
E(n) = sum(sum(P(:,:,n).^2)); %%能量
for i = 1:16
for j = 1:16
if P(i,j,n)~=0
H(n) = -P(i,j,n)*log(P(i,j,n))+H(n); %%熵
end
I(n) = (i-j)^2*P(i,j,n)+I(n); %%惯性矩
Ux(n) = i*P(i,j,n)+Ux(n); %相关性中μx
Uy(n) = j*P(i,j,n)+Uy(n); %相关性中μy
end
end
end
for n = 1:4
for i = 1:16
for j = 1:16
deltaX(n) = (i-Ux(n))^2*P(i,j,n)+deltaX(n); %相关性中σx
deltaY(n) = (j-Uy(n))^2*P(i,j,n)+deltaY(n); %相关性中σy
C(n) = i*j*P(i,j,n)+C(n);
end
end
C(n) = (C(n)-Ux(n)*Uy(n))/deltaX(n)/deltaY(n); %相关性
end
%--------------------------------------------------------------------------
%求能量、熵、惯性矩、相关的均值和标准差作为最终8维纹理特征
%--------------------------------------------------------------------------
a1 = mean(E)
b1 = sqrt(cov(E))
a2 = mean(H)
b2 = sqrt(cov(H))
a3 = mean(I)
b3 = sqrt(cov(I))
a4 = mean(C)
b4 = sqrt(cov(C))
sprintf('0,45,90,135方向上的能量依次为: %f, %f, %f, %f',E(1),E(2),E(3),E(4)) % 输出数据;
sprintf('0,45,90,135方向上的熵依次为: %f, %f, %f, %f',H(1),H(2),H(3),H(4)) % 输出数据;
sprintf('0,45,90,135方向上的惯性矩依次为: %f, %f, %f, %f',I(1),I(2),I(3),I(4)) % 输出数据;
sprintf('0,45,90,135方向上的相关性依次为: %f, %f, %f, %f',C(1),C(2),C(3),C(4)) % 输出数据;
I=imread(‘image.jpg’);I=rgb2gray(I);imshow(I);
disp(I);
I=imresize(I,1/40);
[M,N]=size(I);
for i=1:M
for j=1:N
for n=1:16
if(n-1)*16<=I(i,j)&I(i,j)<=(n-1)*16+15
I(i,j)=n-1;
end
end
end
end
P=zeros(16,16,4);
for m=1:16
for n=1:16
for i=1:M
for j=1:N
if j<N & I(i,j)==m-1 & I(i,j+1)==n-1
P(m,n,1)=P(m,n,1)+1;
end
if i>1&j<N&I(i,j)==m-1 & I(i-1,j+1)==n-1
P(m,n,2)=P(m,n,2)+1;
end
if i<M & I(i,j)==m-1 & I(i+1,j)==n-1
P(m,n,3)=P(m,n,3)+1;
end
if i<M &j<N & I(i,j)==m-1 & I(i+1,j+1)==n-1
P(m,n,4)=P(m,n,4)+1;
end
end
end
end
end
for n=1:4
P(:,:,n)=P(:,:,n)/sum(sum(P(:,:,n)));
end
H=zeros(1,4);
CON=H;
Ux=H;Uy=H;
deltaX=H; deltaY=H;
COR=H;
L=H;
for n=1:4
ASM(n)=sum(sum(P(:,:,n).^2));
for i=1:16
for j=1:16
if P(i,j,n)~=0
H(n)=-P(i,j,n)*log(P(i,j,n))+H(n);
end
CON(n)=(i-j)^2*P(i,j,n)+CON(n);
Ux(n)=i*P(i,j,n)+Ux(n);
Uy(n)=j*P(i,j,n)+Ux(n);
end
end
end
for n=1:4
for i=1:16
for j=1:16
deltaX(n)=(i-Ux(n))^2*P(i,j,n)+deltaX(n);
deltaY(n)=(j-Uy(n))^2*P(i,j,n)+deltaY(n);
COR(n)=i*j*P(i,j,n)+COR(n);
L(n)=P(i,j,n)^2/(1+(i-j)^2)+L(n);
end
end
COR(n)=(COR(n)-Ux(n)*Uy(n))/deltaX(n)/deltaY(n);
end
sprintf('0,45,90,135 方向上的能量依次为: %f,%f,%f,%f',ASM(1),ASM(2),ASM(3),ASM(4))
sprintf('0,45,90,135 方向上的熵依次为: %f,%f,%f,%f',H(1),H(2),H(3),H(4))
sprintf('0,45,90,135 方向上的惯性矩依次为: %f,%f,%f,%f',CON(1),CON(2),CON(3),CON(4))
sprintf('0,45,90,135 方向上的相关性依次为: %f,%f,%f,%f',COR(1),COR(2),COR(3),COR(4))
sprintf('0,45,90,135 方向上的逆差距依次为: %f,%f,%f,%f',L(1),L(2),L(3),L(4))
a1=mean(ASM);b1=sqrt(cov(ASM));
a2=mean(H);b2=sqrt(cov(H));
a3=mean(CON);b3=sqrt(cov(CON));
a4=mean(COR);b4=sqrt(cov(COR));
a5=mean(L);b5=sqrt(cov(L));
sprintf('能量的均值和标准差分别为: %f,%f',a1,b1)
sprintf('熵的均值和标准差分别为: %f,%f',a2,b2)
sprintf('惯性矩的均值和标准差分别为: %f,%f',a3,b3)
sprintf('相关性的均值和标准差分别为: %f,%f',a4,b4)
sprintf('逆差距的均值和标准差分别为: %f,%f',a5,b5)