matlab.zip_shelterf5v_图像分割 小波_小波图像分割

%----------- Input ---------------- Orig_I1=rgb2gray(imread('ss.JPG')); Orig_I = double(Orig_I1) / 256; figure,imshow(Orig_I);title('原来图像');axis on; %----------- Pre-processing ---------------- now=[]; [nRow, nColumn] = size(Orig_I); L=16; %the size of moving windows %expand_image=zeros(2*nRow,2*nColumn); b=flipdim(Orig_I,1); c=[b;Orig_I;b]; d=flipdim(c,2); expand_image=[d c d]; seg_image=zeros(L); [Row, Column] = size(expand_image); %----------- segmenting the image ---------------- for i=nRow+1:Row-nRow for j=nColumn+1:Column-nColumn x=i;y=j; disp(['--> discompose the point at row ' num2str(x-nRow) ' column ' num2str(y-nColumn) '.']); for m=1:L for n=1:L seg_image(m,n)=expand_image(x-L/2+1,y-L/2+1); y=y+1; end x=x+1; y=j; end %----------- Wavelet Decomposition ---------------- %n = size(seg_image,1); %n_log = log2(n); level = 2; type = 'bior4.4'; [Lo_D,Hi_D,Lo_R,Hi_R] = wfilters(type); [E_W, S] = func_DWT(seg_image, level, Lo_D, Hi_D); E_W=E_W'; now=[now;E_W]; end end n_now=now; %normalize the now matrix [now_row,now_column]=size(now); max_now=max(now); min_now=min(now); fm=max_now-min_now; for j=1:now_column for i=1:now_row now(i,j)=( now(i,j)-min_now(j) )/ fm(j); end end % FCM method making sort disp('FCM method making sort.'); [center,U,obj_fcn] = fcm(now,4); maxU = max(U); index1 = find(U(1,:) == maxU); % Find the data points with highest grade of membership in cluster 1 index2 = find(U(2,:) == maxU); % Find the data points with highest grade of membership in cluster 2 index3 = find(U(3,:) == maxU); % Find the data points with highest grade of membership in cluster 3 for r=1:nRow for c=1:nColumn disp(['--> fcm making sorts at row ' num2str(r) 'column ' num2str(c) '.']); m=(r-1)*nColumn+c; size1=size(index1,2); size2=size(index2,2); size3=size(index3,2); q=1; size_now=max(max(size1,size2),size3); while (q<size_now+1) if q<size1+1 if index1(q)==m Orig_I1(r,c)=255; break; end end if q<size2+1 if index2(q)==m Orig_I1(r,c)=128; break; end end if q<size3+1 if index3(q)==m Orig_I1(r,c)=64; break; end end q=q+1; end if(q==size_now+1) Orig_I1(r,c)=0; end end end figure,imshow(Orig_I1);title('分割后图像');axis on;