试用软件

how to calculate actual snr of image in matlab?

151 次查看（过去 30 天）

显示更早的评论

vipul utsav 2012-12-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/55649-how-to-calculate-actual-snr-of-image-in-matlab

评论： Alan Keenan 2021-9-20

采纳的回答： Image Analyst

how to calculate actual snr of image in matlab?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Image Analyst 2012-12-4

3
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/55649-how-to-calculate-actual-snr-of-image-in-matlab#answer_67448

Get the signal - that's your "true" noiseless image.

Get the noise - that's your actual noisy image minus the "true" noiseless image.

Divide them element by element, then take the mean over the whole image.

12 个评论
显示 10更早的评论隐藏 10更早的评论

Image Analyst 2012-12-5

在 MATLAB Online 中打开

NO, that's certainly not true. In fact it's the opposite - it's what you said not to do. What if the mean of the noise was zero? Your method would say that the SNR is infinite even though there's substantial noise. The ratio of the means IS NOT the same as the mean of the ratios. Run this code to see why:

clc;    % Clear the command window.
close all;  % Close all figures (except those of imtool.)
clear;  % Erase all existing variables.
workspace;  % Make sure the workspace panel is showing.
format longg;
format compact;
fontSize = 20;
% Create noise-free image.
noiseFreeImage = 100*ones(240, 320);
subplot(1,2,1);
imshow(noiseFreeImage, []);
% Add +/- 10 gray levels of noise to it
noisyImage = noiseFreeImage + 20*rand(size(noiseFreeImage))-10;
subplot(1,2,2);
imshow(noisyImage, []);
% Get the noise-only image
noiseOnlyImage = abs(noisyImage - noiseFreeImage);
% Calculate the mean of each
signalMean = mean(noiseFreeImage(:))
noiseOriginal = mean(noisyImage(:))
noiseOnlyMean = mean(noiseOnlyImage(:))
% Get ratio of the means
snr1 = signalMean/noiseOnlyMean
% Do element by element
snrImage = signalMean ./ noiseOnlyImage;
% Get the mean of the SNR image.
snr2 = mean(snrImage(:))

Image Analyst 2012-12-5

Well, what if your signal is 100 everywhere, and your noisy image is 100 plus or minus 100, so it goes from 0 to 200. Don't you agree that that is a very very noisy image and it should have a low signal to noise ratio? It's total noise - all over the place. But it's mean is still 100 - same as your signal. So just looking at the means, you can't tell the difference between your perfect noise-free signal and the incredibly noisy signal. How is that any good?

By the way, you can always omit pixels where the noise is zero if you want.

vipul utsav 2012-12-5

thanks for your valuable guidance

请先登录，再进行评论。

更多回答（1 个）

Dhurgham Al-karawi 2017-8-29

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/55649-how-to-calculate-actual-snr-of-image-in-matlab#answer_279533

编辑：Image Analyst 2017-8-29

在 MATLAB Online 中打开

Hi,

Here you are.

img=imread('.JPG');
img=double(img(:));
ima=max(img(:));
imi=min(img(:));
ims=std(img(:));
snr=10*log((ima-imi)./ims);

4 个评论
显示 2更早的评论隐藏 2更早的评论

Image Analyst 2019-10-17

A signal to noise ratio requires both a signal, and noise. If you don't have each of those, you can't do it. However there are single image quality metrics. Try this link.

Alan Keenan 2021-9-20

Perform a histogram on your image and then decide at what level you consider to be signal, i.e., above a value is signal and below that value is noise. If your image is in dB then you might use -6dB as the threshold value. In my case my max signal value is always 0dB as it has been normalised to the max value and converted to dB. So you can say anything less than -6dB is noise and anything greater than -6dB is signal. An easy way to calculate the noise floor is rms(rms(image));. This means that if your signal has a max value of 0dB then the peak SNR for your signal = 0dB + the value of the rms calculation.

请先登录，再进行评论。

请先登录，再回答此问题。

类别

Image Processing and Computer Vision Image Processing Toolbox

在 Help Center 和 File Exchange 中查找有关 Image Processing Toolbox 的更多信息

标签

产品

Image Processing Toolbox

Community Treasure Hunt

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

Translated by

试用软件