最大熵图像恢复

说明

图像降质的过程通常可以用以下模型描述:

$$Y=H\ast X+N$$

Y表示观测到的图像，H表示点扩展函数(Point Spread Function, PSF)。“*“表示二维卷积运算。

参考陈春涛 ,黄步根发表的文献

使用以下公式进行迭代计算:
$$X_k=cexp({\lambda }_k\ast H)$$
$${\lambda }_{k+1}={\lambda }_k+\Delta {\lambda }_k$$
$$\Delta {\lambda }_k=log(Y)-log(H\ast X_k)$$

matlab代码

clc,close all
clear

I = im2double(imread('cameraman.tif'));
figure
imshow(I)
title('Original Image');
saveas(gcf,'save1.bmp')

% add blur
LEN = 21;
THETA = 11;
PSF = fspecial('motion', LEN, THETA);
blurred = imfilter(I, PSF, 'conv', 'circular');

% add noise
noise_mean = 0;
noise_var = 0.0001;
blurred_noisy = imnoise(blurred, 'gaussian', ...
                        noise_mean, noise_var);     
                                  
figure, imshow(blurred_noisy)
title('Simulate Blur and Noise')
saveas(gcf,'save2.bmp')

% MEM method

lambda = ones(size(blurred_noisy));
I2 = ones(size(blurred_noisy));
c=0.4;

for i=1:20
    dlambda = (log(blurred_noisy) - log(imfilter(I2, PSF, 'conv', 'circular')));
    lambda = lambda + dlambda;
    I2 = c*exp(imfilter(lambda, PSF, 'conv', 'circular'));
end

figure, imshow(I2)
title('MEM method')
saveas(gcf,'save.bmp')

效果

参考文献

[1] 陈春涛, 黄步根. 模糊图像的最大熵恢复[J]. 中国司法鉴定, 2010(01):76-78.