Is there any conventional approach to calculate the following integral:

a (x,y) = int [ A (u,v) * H (x,y,u,v) * exp [ i (x*u + y*v) ] dudv] 

The part A(u,v)*exp[ i (x*u + y*v)] represents a conventional FFT. But I also have a set of functions Hx,y(u,v) which have their own distributions in u,v - plane.

One direct way is to perform N calculations substituting Hi,j function one by another. The question is if there is some general FFT approach optimizing the calculus by reducing the number of calls?

More Georgy Sergeevich Kalenkov's questions See All
Similar questions and discussions