# FFT of an image

426 views (last 30 days)
Rachel on 18 Jun 2012
Hello,
I am a new Matlab user trying to compute the FFT of a set of images using the following code:
F=fft2(double(I));
S=fftshift(F);
L-log2(S);
A=abs(L);
imagesc(A)
The code works fine and produces a 2D power spectrum of my image and during subsequent processing I histogram the data in A according to distance form the centre. I have two questions:
1. How do I read this output? The majority of the data cluster around one point on the x-axis...is the corresponding value on the x-axis in cycles/deg? If not how do I get this information from the analysis?
2. How do I read in multiple images at a time so that I don't have to do this analysis for each individual image?
Thank you so much for your help!!
Rachel
##### 2 CommentsShowHide 1 older comment
Dr. Seis on 18 Jun 2012
doing "fftshift" twice is not necessary:
>> a = rand(4,4)
a =
0.4669 0.3845 0.6051 0.5502
0.0665 0.6463 0.9255 0.3866
0.3338 0.2953 0.9869 0.0707
0.7681 0.4264 0.7705 0.7959
>> fftshift(a)
ans =
0.9869 0.0707 0.3338 0.2953
0.7705 0.7959 0.7681 0.4264
0.6051 0.5502 0.4669 0.3845
0.9255 0.3866 0.0665 0.6463

Dr. Seis on 18 Jun 2012
See my answer here for a coded up description of what the 2D Fourier transform does using a random 2D image as an example (should help with your first question):
As to your second question, why wouldn't you do the analysis on an image by image basis? Or is it you want to read in multiple images at one time and then perform some operation on each of them in a similar way?