CLAHE2_CLAHE_信息熵；matlab_自适应_

image = imread('C:\Users\Administrator\Desktop\1212.png'); %image = rgb2gray(image); edgeThreshold = 0.2; amount = 0.5; data = localcontrast(image, edgeThreshold, amount); PSNR1 = psnr(data, image); fprintf('\n The Peak-SNR value of LCM is %0.4f', PSNR1); t=1; % start index of the window (row) limit=8; % window size of the contextual area endt=limit; % end index of the window (row) eqdata=zeros(size(data,1),size(data,2)); for x=1:size(data,1) q=1; % start index of the window (column) endq=limit; % end index of the window (column) %% TO move Window to right and bottom, after exceeding the limit for y=1:size(data,2) eqdata(x,y)=0; if (x>t+limit-1) %to move rows of window right t=t+limit; endt=limit+t-1; end if (y>q+limit-1) %to move cols of window right q=q+limit; endq=limit+q-1; end if (endt>size(data,1)) %to shift the rows of window to the next line % t=t-64; endt=size(data,1); end if (endq>size(data,2)) %to shift the cols of window to the next line % q=q-64; endq=size(data,2); end %% Counting the number of pixels in each contextual area for i=t:endt for j=q:endq if data(x,y)==data(i,j) eqdata(x,y)=eqdata(x,y)+1; %storing count of pixels values in matrix eqdata end end end end end %% Second Pass: Calculate partial rank, redistributed area and output values. output=zeros(size(data,1),size(data,2)); cliplimit=0.3; % Cliplimit can vary between 0 to 1. t=1; endt=limit; for x=1:size(data,1) q=1; endq=limit; %% TO move Window to right and bottom, after exceeding the limit for y=1:size(data,2) cliptotal=0; partialrank=0; if (x>t+limit-1) t=t+limit; endt=limit+t-1; end if (y>q+limit-1) q=q+limit; endq=limit+q-1; end if (endt>size(data,1)) % t=t-64; endt=size(data,1); end if (endq>size(data,2)) % q=q-64; endq=size(data,2); end %% For each pixel (x,y), compare with cliplimit and accordingly do the clipping. Calculate partialrank. for i=t:endt for j=q:endq if eqdata(i,j)>cliplimit incr=cliplimit/eqdata(i,j); else incr=1; end cliptotal=cliptotal+(1-incr); if data(x,y)>data(i,j) partialrank=partialrank+incr; end end end %% New distributed pixel values can be found from redistr and will be incremented by partial rank. redistr=(cliptotal/(limit*limit)).*data(x,y); output(x,y)=partialrank+redistr; end end figure('name','LCM-CLAHE','NumberTitle','off') %da=adapthisteq(data); imshow([image uint8(output)]); % Concatenate original and CLAHE image figure('name','Histogram of input image','NumberTitle','off'); imhist(image); %Show the histogram of original image, LCM output and CLAHE output figure('name','Histogram of LCM output','NumberTitle','off');imhist(data); figure('name','Histogram of CLAHE output','NumberTitle','off');imhist(uint8(output)); PSNR = psnr(uint8(output), image); fprintf('\n The Peak-SNR value of LCM-CLAHE is %0.4f', PSNR); normal = adapthisteq(image,'clipLimit',0.02,'Distribution','rayleigh'); %imshow([image normal]); PSNR0 = psnr(normal,image); fprintf('\n The Peak-SNR value of CLAHE is %0.4f', PSNR0); he = histeq(image); PSNR2 = psnr(he,image); fprintf('\n The Peak-SNR value of HE is %0.4f', PSNR2);