2d fft for many images

If I would like to compute and plot the 2D frequency spectrum for multiple images (50 frames), can I do that for each individual image and the avrege them?

here is the code:

min=zeros;

for K = 1 :1500:75000

this_file = strcat('C:\Users\Admin\Desktop\Tutorial\frame',num2str(K),'.tif');

im_input= rgb2gray(imread(this_file));

z_min=10*log10(abs(fftshift(fft2(im_input))));

min=min+z_min;

end

PSD_min= (min)/ 50;

______________

1- is this a right way compute the spectrum?

2- If I would like to remove the DC component, I tried

z_min=10*log10(abs(fftshift(fft2(im_input-mean2((im_input))... do you recommend other way?

Thanks

[Jonas]

it looks like the correct way to conpute the 2d spectrum, also the removal of the dc component ist correct. the questions if it is ok to add up the spectra depends on your goal, like already wrote you will get an anverage spectrum, this would make for me only sense, if the images are similar, otherwise you add up information that do not necessarily belong to each other / are related to each other.

you could also consider to add up the complex spectra (without the abs() ) and afterwards you can use abs on the summed spectrum and average it

[Image Analyst]

@Jonas Looks like a fine "Answer" to me so it should be down below in the "Answer" section.

Answer 1

[Jonas]

it looks like the correct way to conpute the 2d spectrum, also the removal of the dc component is correct. the questions if it is ok to add up the spectra depends on your goal, like you already wrote you will get an anverage spectrum, this would make for me only sense, if the images are similar, otherwise you add up information that do not necessarily belong to each other / are related to each other. you could also consider to add up the complex spectra (without the abs() ) and afterwards you can use abs on the summed spectrum and average it

[omar Alharbi]

Thanks for your feedback. Yse the images are similar.

I appreciate your comments

Answer 2

[Sulaymon Eshkabilov]

To remove the DC from your calculated values, you can use:

DC=mean2(im_input);
IM2=im_input-DC;