Can any one give solution for removing ripples in below image

Question

Sivakrishna 2012-9-17

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/48399-can-any-one-give-solution-for-removing-ripples-in-below-image

采纳的回答： Image Analyst

https://www.dropbox.com/s/2bsf32que8fz77w/ripplesimg.jpg

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2012-9-17

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/48399-can-any-one-give-solution-for-removing-ripples-in-below-image#answer_59127

That looks like a combination of errors. The first problem looks like a demosaicing error. Then it underwent blurring. Try deblurring with blind deconvolution (deconvblind) or richardson lucy (deconvlucy) (Examples for both in the help), then try some ad hoc method to fix the demosaicing problem. Report back here with your code if you have problems.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Sivakrishna 2012-9-17

编辑：Walter Roberson 2012-9-18

在 MATLAB Online 中打开

hi this is my deconvolution code..

function[deconv,psf]=blind_dconv(z,iterations,hesize,alpha,fe,k)
% clc,clear all
% close all
% a0=double(imread('cameraman.tif'));
% h = fspecial('motion',1);
% 
% z = imfilter(a0,h,'replicate');
% figure,imshow(z,[])
% iterations=1;
% hesize=1;
% alpha=0.01;
% k=0.01;
% fe=10;
% if(mod(hesize,2)~=1)
%   error('Size of psf must be odd');
% end
YSIZE=size(z,1);
XSIZE=size(z,2);
SIZEMAX=max([XSIZE YSIZE]);
%z is the linearly degraded image, fe is the image estimate, and k is the
%finite support. he is the PSF estimate.
%This is the PSF ESTIMATE (we estimate it as an impulse)
he=zeros(1,hesize);
he(ceil(hesize/2))=1;
he=he'*he;
G=fft2(z,(2*SIZEMAX),(2*SIZEMAX)); %z is the degraded image
n=0;
while(n<iterations)
  Fe=fft2(fe,(2*SIZEMAX),(2*SIZEMAX));Hoe=fft2(he,(2*SIZEMAX),(2*SIZEMAX));
    %impose fourier constraints
    Hne=(G.*conj(Fe))./(abs(Fe).^2+(alpha./(abs(Hoe).^2)));
    hne=ifft2(Hne);
    %impose blur constraints
    hne=abs(hne(1:hesize,1:hesize));
    Hoe=fft2(hne,(2*SIZEMAX),(2*SIZEMAX));
    %Impose Fourier constraints
    Fne=(G.*conj(Hoe))./((abs(Hoe).^2+(alpha./abs(Fe).^2)));
    fne=ifft2(Fne);
    %impose image constraints
    fne=abs(fne(1:YSIZE,1:XSIZE));
    fne(~k)=0;
    fe=abs(fne);he=abs(hne);
    n=n+1;
  end
deconv=fne;
psf=hne;