如何实现高原极限直方图均衡算法

Question

我一直在尝试实现一个简单而简短的算法。我将结果的直方图与原始结果的直方图进行比较，总是有很大的不同。

我一次又一次地查看我的代码。但我看不出有什么问题。我发布了步骤和我的尝试。我的错在哪里？

histogram_of_image = imhist(input_image);
modified_histogram = zeros(1,256);
modified_histogram(1:256)  = (log(histogram_of_image(1:256)+(a))).^(beta);

these_elements_are_not_zero = modified_histogram~= 0;
sum2 = sum(modified_histogram(these_elements_are_not_zero));
cnt2 = size(these_elements_are_not_zero,2);
tcl = sum2/cnt2;
clipped_histogram = zeros(1,256);
for i=1:256
    if((modified_histogram(i)) >= tcl)
    clipped_histogram(i) = tcl;    
    else 
    clipped_histogram(i) = (modified_histogram(i));   
    end
end

PDa = zeros(1,256);
PDa = clipped_histogram / (sum(clipped_histogram)); 
CDa = zeros(1,256);  %create CDa in formula
CDa(1) = PDa(1) ;
for i=2:256
     CDa(i) =  PDa(i) +  CDa(i-1);  
end
value_after_enhancement =  zeros(1,256); 
value_after_enhancement = (255 * CDa); 
b=uint8(0);
output_image=zeros(width,height);
output_image =   value_after_enhancement(input_image+1);
figure;imshow(uint8(output_image));
original_result=imread('testsonuc.bmp');
original_result=double(rgb2gray(original_result));
image_of_dif= zeros(width,height);
figure;imshow(uint8(original_result));

Answer 1

除了绘制结果的一个小细节之外，一切都是正确的，您应该将输出图像从双精度整数转换为无符号整数。output_image = uint8(output_image);

%value_after_enhancement =  zeros(1,256);
value_after_enhancement = (255 * CDa)+1;
%b = uint8(0);
%output_image = zeros(width,height);
output_image = value_after_enhancement(input_image);
output_image = uint8(output_image);
%dif_image = output_image ./ input_image;
%===================================================================
figure('Name','Algorithm Result','NumberTitle','off');
subplot(211);
imshow(output_image);
subplot(212);
Out_Hist = imhist(output_image);
plot(Out_Hist);