Adding Impulse Noise in Image manually

Question

How can salt & pepper noise be added to an image separately depending on the PROBABILITY of salt and pepper. imnoise would take ONE value for the noise density as a whole and that value is the measure for both salt (white dots) and pepper (black dots). I want to know if we have just to add white (salt) and then black (pepper) noise separately with two different probabilities. What equation would have to be used?

Answer 1

clc;
close all;
originalImage = imread('Cameraman.tif');
[rows cols] = size(originalImage);
totalPixels = int32(rows * cols);
subplot(1, 2, 1);
imshow(originalImage);
percentage = str2double(cell2mat(inputdlg('Enter the percent noise: ', 'Enter answer', 1, {'2'}))) / 100.;
numberOfNoisePixels = int32(percentage * double(rows) * double(cols));
locations = randi(totalPixels, [numberOfNoisePixels, 1]);
noisyImage = originalImage;
noisyImage(locations) = 255;
subplot(1, 2, 2);
imshow(noisyImage, []);

Source

https://groups.google.com/forum/#!topic/comp.soft-sys.matlab/YcF2xZwnq1o

That does Salt Noise, Pepper noise would be

noisyImage(locations) = 0;

instead of

noisyImage(locations) = 255;

Answer 2

img = .5*ones(100,200); %// example image
p_salt = .05; %// probability of salt
p_pepper = .01; %// probability of pepper

if strcmp(class(img),'uint8')
    salt_value = uint8(255);
else
    salt_value = 1;
end
pepper_value = 0;

aux = rand(size(img)); %// generate random values
img(aux<=p_salt) = salt_value;  %// add salt
img((aux>p_salt) & (aux<=p_salt+p_pepper)) = pepper_value; %// add pepper

imshow(img) %// show image

This approach is similar to that used in imnoise, and avoids salt and pepper to be added at the same pixels. It assumes p_salt + p_pepper is at most 1.

Answer 3

This code is simple and useful(Same imnoise in MATLAB)

im=imread('Parrot.jpg');
B=rgb2gray(im);
%if Pa==Pb;
percen=10;
%Noise level 10
Prob_den_f=255*percen/100;
NoiseImg = B;
Rmatrix = randint(size(B,1),size(B,2),[0,255]);
NoiseImg(Rmatrix <=Prob_den_f/2) = 0;
NoiseImg(Rmatrix >Prob_den_f/2&Rmatrix<Prob_den_f) =255;
subplot(1,2,2),imshow(NoiseImg),title('Add ''Salt and Pepper'' Noise'); 
subplot(1,2,1),imshow(B),title('Original Image');