matlab实现梯度倒数加权滤波

a=imread('lena.bmp'); figure(1); imshow(a); a=imnoise(a,'gaussian',0.005); figure(2); imshow(a); a=double(a); [dep,wide]=size(a); new_image=ones(size(a)); r=0.5; for i=2:dep-1 for j=2:wide-1 g=0; for m=-1:1 for n=-1:1 if(a(i+m,j+n)-a(i,j)==0) g(m+2,n+2)=0; else g(m+2,n+2)=1/abs(a(i+m,j+n)-a(i,j)); end end end G=sum(sum(g)); for m=-1:1 for n=-1:1 w(m+2,n+2)=g(m+2,n+2)/G; end end new_image(i,j)=a(i,j)*r+(w(-1+2,-1+2)*a(i-1,j-1)+w(-1+2,0+2)*a(i-1,j)+w(-1+2,1+2)*a(i-1,j+1)+w(0+2,-1+2)*a(i,j-1)+w(0+2,1+2)*a(i,j+1)+w(1+2,-1+2)*a(i+1,j-1)+w(1+2,0+2)*a(i+1,j)+w(1+2,1+2)*a(i+1,j+1))*(1-r); end end for i=2:dep-1 new_image(i,1)=new_image(i,2); new_image(i,wide)=new_image(i,wide-1); end new_image(1,:)=new_image(2,:); new_image(dep,:)=new_image(dep-1,:); figure(3); imshow(uint8(a)) figure(4); imshow(uint8(new_image)) imwrite(uint8(new_image),'reci_gradu_weighted_0.5.bmp','bmp')