【图像处理】基于Hough变换结合分数阶变分PDE实现图像去雨含Matlab源码-CFANZ编程社区

【图像处理】基于Hough变换结合分数阶变分PDE实现图像去雨含Matlab源码

1 简介

图像处理过程中,保持图像的边缘,纹理等重要结构特征是非常重要的.要保持这重要的图像信息,在图像处理时就要对图像中的卡通,边缘,纹理及噪声等不同成分进行区分和处理,也就需要对图像的不同成分进行合适的数学建模. 本论文结合图像去噪问题的研究,利用小波分析理论与方法,函数空间图像建模理论与方法以及新的数学工具如分数阶导数等,探索对图像中的不同成分的新的建模方法,并在此基础上利用正则化理论和方法,算子分裂法,变分法等提出新的图像去噪变分PDE模型和算法.论文结合图像去噪提出的基于分数阶变分PDE的图像建模理论和方法,可以进一步推广到图像分割,超分辨率重建等其它图像处理领域,具有重要的理论意义和广泛的实际应用前景.

2 部分代码

function [img_out, img_tag] = GeneratorCore(img)
% Find raindrop location, generate mask, masked img
% Get image
I = img;
% Set boundaries
Rmin = 1;
Rmax = 10;
% Find all the bright circles in the image
[centersBright, radiiBright] = imfindcircles(I,[Rmin Rmax], ...
    'ObjectPolarity','bright');
% Find all the dark circles in the image
[centersDark, radiiDark] = imfindcircles(I, [Rmin Rmax], ...
    'ObjectPolarity','dark');
% % Plot bright circles in blue
% viscircles(centersBright, radiiBright.*1.5,'EdgeColor','b');
% % Plot dark circles in dashed red boundaries
% viscircles(centersDark, radiiDark.*1.5,'LineStyle','--');
% figure;
% viscircles(centersBright, radiiBright.*1.5,'EdgeColor','b');
% % Plot dark circles in dashed red boundaries
% viscircles(centersDark, radiiDark.*1.5,'LineStyle','--');
% title('Detection result');
% grid on;
% Normalization
cB = round(centersBright);
rB = round(radiiBright);
cD = round(centersDark);
rD = round(radiiDark);
% Generate Mask
mask = ones(size(I,1),size(I,2));
mask2 = setNAN(mask,[cB;cD],[rB;rD],2);
% figure; subplot(1,2,1);imshow(mask2);
% imwrite(mask2, './mask_B.bmp', 'BMP');
iii = I;
title('Tagged image');
% Apply filtering
for a = 1:size(I,1)
    for b = 1:size(I,2)
        if(mask2(a,b) == 1)
            iii(a,b) = 255;
        end
    end
end
% subplot(1,2,2); imshow(iii)
img_out = uint8(iii);
img_tag = uint8(255*mask2);
grid on;
title('Detect result');
end