You’ve basically outlined the right approach which I’ll summarize here:
1) Do the 2D FFTs.
2) Multiply them by a weighting factor. If you’re FFTs are in (Real, Imaginary) form, multiply both components by the weight, and if (Magnitude, phase) just multiple the magnitude. If you’re interested in just high and low frequencies, this will likely be radial weights, giving weights to things depending on their frequency.
3) Add the two together.
4) Inverse 2D FFT.
You can do the addition (item 3), either before or after the iFFT. Also, as Paul R mentioned, you can do the 2D filtering in the spatial domain instead of the frequency domain.
0
solved How to combine low and high frequencies of two images in Matlab [closed]