有谁matlab写过基于自适应滤波的Retinex图像增强算法

Question

其中的W(x,y)滤波窗函数怎么用matlab描述？

Answer 1

根据最后一个公式计算w(x,y), 实际上就是计算梯度水平(水平和垂直方向两个方向综合)

I = imread('test3.jpg');       %读入图象

subplot(121);imshow(I);

I = double(rgb2gray(I));

[m,n]=size(I);

H1 = [-1 -2 -1

       0  0  0

       1  2  1];

H2 = [-1  0  1

      -2  0  2

      -1  0  1];

w0 = (abs(conv2(I,double(H1),'same')) + abs(conv2(I,double(H1),'same')))/2+ones(m,n);

subplot(122);imshow(uint8(w0));

w = 1./w0;

追问

最后一个问题，它滤波后图像 Io（x,y) 用matlab怎么写

追答

clear all;

I = imread('lena.bmp');       %读入图象

% I = rgb2gray(I);

figure;imshow(I);

I = double(I);

[m,n]=size(I);

H1 = [-1 -2 -1

       0  0  0

       1  2  1];

H2 = [-1  0  1

      -2  0  2

      -1  0  1];

w0 = (abs(conv2(I,double(H1),'same')) + abs(conv2(I,double(H1),'same')))/2+ones(m,n);

w = 1./w0;

%计算N(x,y)

% 设窗口D=7;

D = 7;

N = zeros(m,n);

for x = 1:m-D

    for y = 1:n-D

        N(x,y)=sum(sum(w(x:x+D-1,y:y+D-1)));

    end

end

%计算I0

for x = 1:m-D

    for y = 1:n-D

        I0(x,y)= sum( sum( w(x:x+D-1,y:y+D-1).*I(x:x+D-1,y:y+D-1)) )./N(x,y);

        I0h(x,y)=max(I0(x,y),I(x,y));

    end

end

figure; imshow(uint8(I0h));