【图像重建】基于POCS算法实现超分辨率图像重建，PSNR附Matlab

% 超分辨率图像重建（基于POCS算法）

% 读取低分辨率图像
lowResImage = imread(‘low_resolution_image.jpg’);

% 参数设置
scaleFactor = 2; % 倍数因子
numIterations = 10; % 迭代次数

% 图像尺寸调整
highResSize = scaleFactor * size(lowResImage);
highResImage = imresize(lowResImage, highResSize, ‘bicubic’);

% 初始化超分辨率图像
superResImage = highResImage;

% 迭代重建
for iter = 1:numIterations
% 从超分辨率图像生成低分辨率图像的估计
estimatedLowResImage = imresize(superResImage, size(lowResImage), ‘bicubic’);

% 计算残差图像
residual = double(lowResImage) - double(estimatedLowResImage);

% 从残差图像中恢复细节
detailEnhancedImage = superResImage + residual;

% 对细节增强图像进行尺寸调整
superResImage = imresize(detailEnhancedImage, highResSize, 'bicubic');

end

% 计算PSNR
mse = sum(sum((double(highResImage) - double(superResImage)).^2)) / numel(highResImage);
maxIntensity = double(max(highResImage(😃));
psnr = 10 * log10((maxIntensity^2) / mse);

% 显示结果
figure;
subplot(1, 2, 1);
imshow(highResImage);
title(‘High-Resolution Image’);
subplot(1, 2, 2);
imshow(superResImage);
title(['Super-Resolution Image (PSNR: ‘, num2str(psnr), ’ dB)’]);