How to make salt pepper noise own code

After creating a matrix with the for loop, how can we assign the values 0 and 255 in the picture and add salt and pepper noise?

Answer 1

[Ameer Hamza]

Try this

im = imread('pears.png');

figure;
ax1 = axes();
imshow(im);
title(ax1, 'original');

a = 0.1; % 10% pixels altered
b = 0.5; % 50% percent white pixels among all altered pixels

n = numel(im(:,:,1));
m = fix(a*n);
idx = randperm(n, m);

k = fix(b*m);
idx1 = idx(1:k);
idx2 = idx(k+1:end);

idx1 = idx1' + n.*(0:size(im,3)-1);
idx1 = idx1(:);
idx2 = idx2' + n.*(0:size(im,3)-1);
idx2 = idx2(:);

im(idx1) = 255;
im(idx2) = 0;

figure;
ax2 = axes();
imshow(im);
title(ax2, 'noisy');

[Ali Umur Kucur]

THANK YOU VERY MUCH

[Ameer Hamza]

I am glad to be of help.

[Image Analyst]

Ali, did you try my solution (or even see it below):

noisyImage = imnoise(originalImage,'salt & pepper', 0.05); % Or whatever percentage you want.

It's a lot simpler since it uses the built-in function.

[Ali Umur Kucur]

Thank you for your answer, but I have to write this salt and pepper noise with my own code, not with the ready function.

Answer 2

[David Welling]

An easy way to do this is create a salt and pepper noise image to lay in front of the original image. So you need a way to randomly select pixels to make white. This can easily be done by creating a matrix the same size as your picture, filled with random numbers, and then select a cut off point above which you make pixels white, like this:

floor(rand(1000,1000)+0.01)*255;    %array of 1000x1000, with approximately 1 percent white pixels. this can be adjusted by changing the 0.01 in the equation

[Ali Umur Kucur]

thank you for answer.

Answer 3

[Image Analyst]

The easiest way is to use the built-in imnoise() function:

noisyImage = imnoise(originalImage,'salt & pepper', 0.05); % Or whatever percentage you want.

[Ali Umur Kucur]

Thank you for your answer, but I have to write this salt and pepper noise with my own code, not with the ready function.

[Image Analyst]

Why? It's not labeled as homework. If it is your assignment and you turned in Ameer's code as your own, then you could run into trouble with your teacher and institution (possibly cheating). In the future, tag homework with the homework tag so people don't give you complete solutions that will get you into trouble.

Answer 4

[Mykola Ponomarenko]

function [ima,map] = salt_and_pepper(ima, prob)

% ima - grayscale or color input image; prob - probability of salt&pepper noise (0..1)

[y,x,z]=size(ima);

map=repmat(rand(y,x)<prob, [1 1 z]);

sp=repmat(round(rand(y,x))*255, [1 1 z]);

ima(map)=sp(map);

end

Answer 5

[DGM]

This is the way that MIMT imnoiseFB() does it when in fallback mode. This will replicate the behavior of IPT imnoise(). Note that this works regardless of the class of the input image.

inpict = imread('cameraman.tif');
snpdensity = 0.05; % default for imnoise()/imnoiseFB()

s0 = size(inpict);
noisemap = rand(s0);
outpict = im2double(inpict);
mk1 = noisemap < (snpdensity/2);
outpict(mk1) = 0;
outpict(~mk1 & (noisemap < snpdensity)) = 1;

imshow(outpict)