Reputation:
I have basically this:
I = im2double(imread('lena.png'));
% Add gaussian noise
J = imnoise(I,'gaussian',0,0.05);
Now having both undegraded image I
and noised image J
how can I estimate that the variance used for gaussian noise was 0.05
?
Upvotes: 2
Views: 290
Reputation: 272487
You can obtain the noise waveform via subtraction. You can obtain the sample variance with var()
.
Of course, the sample variance is only an estimate for the true variance.
Upvotes: 2
Reputation: 719
Subtract J from I, square each difference, and take the average. Since you know the mean (zero) you can use the number of pixels as the denominator in your average.
Upvotes: 2