I want to ask why everywhere is used to do edge detection in spatial domain and not in frequency domain with FFT of image and apply HP filter. Thank you.


What you claim is simply not true. One step in edge detection is often a simple high-pass (or band-pass) filter, and those are commonly applied using fast convolution, which is a frequency-domain algorithm.

  • 1
    $\begingroup$ Then why is everywhere in edge detection posted info about Sobel operator, Perwitt and etc? Is it easier to do it in spatial domain with convolution or what ? I cant find obvious reason to do edge detection only in spatial domain. $\endgroup$ – Olo Dec 8 '20 at 19:53
  • $\begingroup$ It's easy to explain. It's not easier to do, necessary. For some things it is. For everything nonlinear, it's usually necessary. $\endgroup$ – Marcus Müller Dec 8 '20 at 19:54
  • $\begingroup$ So for linear transformation it is necessary to do it in spatial domain? and in nonlinear it can be done in both (freq and spatial)? Thx for ans u helped me so much $\endgroup$ – Olo Dec 8 '20 at 19:57
  • $\begingroup$ No, exactly the other way around. If something is linear, and you're using a linear transform like the Fourier transform, it can be done in whichever domain is easier/faster/elegant. Non-linear operations don't "translate". $\endgroup$ – Marcus Müller Dec 8 '20 at 20:02
  • $\begingroup$ I dont understand nevertheless. $\endgroup$ – Olo Dec 8 '20 at 20:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.