How should I remove noise from the image which is marked in red color except the crack lines?

clc;
clear;
close all;

%%step 1
image = imread('D:/canal images/crack.jpg');   
%image = imread('C:/Users/Jainee/Documents/MATLAB/frame60.jpg');
figure
imshow(image);

%%step 2
gray = rgb2gray(image);        
imshow(gray);
title('Gray Scale');

%%step 3
croped = imcrop(gray);  
figure
imshow(croped);
title('Croped Image');

%%step 4
sharpen = imsharpen(croped,'Amount',3);   
figure
imshow(sharpen);
title('Sharped Image');

%%step 5
se = strel('Line',100,90);           %%create a disk-shaped structring element
% bothat = imsubtract(imadd(sharpen,imtophat(sharpen,se)),imbothat(sharpen,se));  %%applying bottom hat method
bothat=imbothat(sharpen,se);
figure
imshow(bothat);
title('Bottom Hat Method');

%%step 6
level = graythresh(bothat);     %%return gary level of image
BW = im2bw(bothat,level);  %%convert into binary 
figure
imshow(BW);
title('Binary Image');

%%step 7
% % s1=size(BW);
% % for i=1:s1(1)
% %     for j=1:s1(2)
% %         bw6(i,j)=1-BW(i,j);
% %     end
% % end
% % figure;
% % imshow(bw6);
% bw6=  BW
bw5 = bwareaopen(BW,80,8);
figure
imshow(bw5)
%%title('Size range: 20-50 pixels')
bwt5=BW-bw5;
figure
imshow(bwt5);
se2=strel('disk',2);
% 
bw6=imclose(bwt5,se2);
% bw16 = imerode(bw6,se2);
figure
imshow(bw6);
bw18=bw6;
si = size(croped);
for i=1:si(1)
    for j=1:si(2)
        if(bw18(i,j)==1)
            new_image(i,j,1)=255;
            new_image(i,j,2)=0;
            new_image(i,j,3)=0;
        else
            new_image(i,j,:)=croped(i,j,:);
        end
    end
end
figure
imshow(new_image);

[John D'Errico]

Please learn to format your code readably next time. I fixed it for you here.

When you paste in the code, select it as a block, then click on the "{} Code" button above the edit window. Your code will now be formatted as code, instead of an unreadable mess.

[Jainee Solanki]

Can you find the solution of my question now?

Answer 1

[Image Analyst]

I'd use regionfill() to fill in the red regions with the surrounding perimeter, kind of like it's smeared inward.