Power density spectrum of an image needed for further calculations

5 views (last 30 days)
Hi there,
I need the density spectrum of an image. I found a calculation which looks like this:
img1 = imread('00001.tif');
F = fftshift(fft2(img1));
F_abs = abs(F).^2;
But I want a plain number which represents the whole density for further calculations. Is it mathematecally correct if I take the mean2(F_abs) for this or do I have to integrate something again?
Best wishes, Thorsten

Answers (1)

Sebastian K
Sebastian K on 22 May 2017
Edited: Sebastian K on 22 May 2017
Hi Thorsten,
I believe you need to clarify what type of a "metric" or "measure" you are looking for in order to describe your image. When you say that you are looking for a "plain number", do you mean a scalar value?
"Power spectral density" is defined as the power content of a signal across its spectrum (frequency contents). For instance if we are talking about a simple signal in time domain, the power spectral density will be an array that shows how much power is associated with certain frequencies in the signal.
In the case of an image, you can think of it as a two-dimensional signal. Again, the power spectral density will be a function of the frequencies, and it will be represented as a 2-D array.
If you are looking for a scalar quantity, you should clarify what it is that you are trying to measure.
For instance, if we are talking about a 1-D sequence, signal to noise ratio (SNR) is a scalar quantity that can be used.
For images, you could also look into SNR and Peak SNR , but I am not sure if this is the kind of metric you would want to use. Also note that you would need to provide a reference image in order to calculate the SNR.
I hope this helps.
Sebastian
  1 Comment
Babu Biswas
Babu Biswas on 20 Feb 2020
will u write a code for me to calculate SNR, PSNR, MSE, RMSE, NRMSE for a noisy signal and a noise free signal?
It will be great help for me sir

Sign in to comment.

Community Treasure Hunt

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

Start Hunting!