Fast Fourier Transform for frequency domain

I have some thermograph images that I took from an object every 3 secs. I used the thermography technique to detect some subsurface defetcs. I found a paper which followed the same approach I used in my research ( Attached below). The researchers of this paper had the same problem as I have in my research. The authors couldnot detect some of the defects from the original images but after a simple image processing the authors of this paper could observe the other defects. please look at Figure 3 in this paper. I want to know how to return the fft of an image on a specif frequency? Would you please help me write a code to do the samething for my images?

Answer 1

[Asvin Kumar]

The attached paper says in Section 2.2.1:

Fourier transform (FT) is particularly interesting among the data processing methods as it allows for a retrieval phase and amplitude data from temperature-time history of each pixel.

It looks to me like FFT is applied across time at each pixel separately. The syntax of fft you're looking for is Y = fft(X, n, dim).

The output Y will be complex valued. You should be able to retrieve both the phase and magnitude for the desired frequency from Y. Here's an example usage off fft which plots just the magnitude: https://www.mathworks.com/help/matlab/ref/fft.html#buuutyt-10