matlab-基于滑动窗口滤波模板的高斯滤波实现图像的保边去噪仿真-源码

function result=SGAU(im, radius, iteration,sigma) % 基于滑动窗口滤波模板的高斯滤波 r = radius; %the radius of the side window k = ones(2*r+1,1)/(2*r+1); %separable kernel k_L=k; k_L(r+2:end)=0; k_L = k_L/sum(k_L); %half kernel k_R=flipud(k_L); m = size(im,1)+2*r; n = size(im,2)+2*r; total = m*n; [row, col]=ndgrid(1:m,1:n); offset = row + m*(col-1) - total; im = double(im); result = im; d = zeros(m,n,8,'single'); for ch=1:size(im,3) U = padarray(im(:,:,ch),[r,r],'replicate'); for i = 1:iteration %all projection distances gausFilter = fspecial('gaussian',[3 3],sigma); xy1=[1,1,0;1,1,0;0,0,0]; xy1=gausFilter.*xy1; Y1=xy1./sum(sum(xy1)); % Y1=mapminmax(xy1,0,1); % xy2=[0,1,1;0,1,1;0,0,0]; xy2=gausFilter.*xy2; % Y2=mapminmax(xy2,0,1); Y2=xy2./sum(sum(xy2)); % xy3=[0,0,0;0,1,1;0,1,1]; xy3=gausFilter.*xy3; % Y3=mapminmax(xy3,0,1); Y3=xy3./sum(sum(xy3)); % xy4=[0,0,0;1,1,0;1,1,0]; xy4=gausFilter.*xy4; % Y4=mapminmax(xy4,0,1); Y4=xy4./sum(sum(xy4)); % xy5=[1,1,1;1,1,1;0,0,0]; xy5=gausFilter.*xy5; % Y5=mapminmax(xy5,0,1); Y5=xy5./sum(sum(xy5)); % xy6=[0,0,0;1,1,1;1,1,1]; xy6=gausFilter.*xy6; % Y6=mapminmax(xy6,0,1); Y6=xy6./sum(sum(xy6)); % xy7=[1,1,0;1,1,0;1,1,0]; xy7=gausFilter.*xy7; % Y7=mapminmax(xy7,0,1); Y7=xy7./sum(sum(xy7)); % xy8=[0,1,1;0,1,1;0,1,1]; xy8=gausFilter.*xy8; % Y8=mapminmax(xy8,0,1); Y8=xy8./sum(sum(xy8)); d(:,:,8)=imfilter(U,Y8,'replicate')-U; d(:,:,7)=imfilter(U,Y7,'replicate')-U; d(:,:,6)=imfilter(U,Y6,'replicate')-U; d(:,:,5)=imfilter(U,Y5,'replicate')-U; d(:,:,4)=imfilter(U,Y4,'replicate')-U; d(:,:,3)=imfilter(U,Y3,'replicate')-U; d(:,:,2)=imfilter(U,Y2,'replicate')-U; d(:,:,1)=imfilter(U,Y1,'replicate')-U; %find the minimal signed distance tmp = abs(d); [~,ind] = min(tmp,[],3); index = offset+total*ind; dm = d(index); %signed minimal distance %update U = U + dm; end result(:,:,ch) = U(r+1:end-r,r+1:end-r); end