how to calculate signal to noise ratio for a particular section in an image automatically

1 Ansicht (letzte 30 Tage)
hi how to calculate signal to noise ratio for a particular section in an image automatically
  9 Kommentare
Image Analyst
Image Analyst am 24 Jun. 2014
Bearbeitet: Image Analyst am 24 Jun. 2014
I don't see a tank. A water tank? A military tank? Anyway, did you try the formula I gave you already a few comments ago? When I asked you "Can you estimate the noise from some area that is supposed to be uniform" and you replied "Yes", what value did you get? Or did you mean that Image Analyst can do it, but seshi can't?
seshi reddy
seshi reddy am 24 Jun. 2014
hey its a military tank having the white spots and the black part in the image is field, actually i did not get you at that time i calculated peak signal to noise ratio
clear;
clc;
load hb04039;
load hb06158;
oi=hb04039;
A = imnoise(oi,'salt & pepper', 0.02);
%Calculate the PSNR.
[peaksnr, snr] = psnr(A, oi);
fprintf('\n The Peak-SNR value is %0.4f', peaksnr);
fprintf('\n The SNR value is %0.4f \n', snr);
by using the above code
The Peak-SNR value is 20.4011
The SNR value is -0.1369
i got the above values but what i want is the signal to noise ratio of the black part in that image

Melden Sie sich an, um zu kommentieren.

Akzeptierte Antwort

Image Analyst
Image Analyst am 24 Jun. 2014
You can get a mask of the tank by thresholding at some value, like 40 or 100 or whatever works well
mask = oi < 40;
% Get the pixels in the mask in the oi image
signalPixels = oi(mask);
% Get the pixels in the mask in the noisy image.
noisePixels = A(mask);
% Get the mean SNR over all those pixels
% First get the SNR on a pixel by pixel basis, then take the mean.
snr = mean(signalPixels ./ noisePixels);
Of course you need to handle zero signal and zero noise cases.
  5 Kommentare
Image Analyst
Image Analyst am 25 Jun. 2014
You can get rid of the nans. See the code below:
A = [1 2 5 6 7 8 nan 20 21 22 23 nan -2 -1 0 nan 1 2]
nanIndexes = isnan(A)
A(nanIndexes) = [] % Remove nans.

Melden Sie sich an, um zu kommentieren.

Weitere Antworten (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by