May1-12, 05:52 AM
I am attempting to implement an algorithm in which I am multiplying an f(x,y) times a (p^2 + q^2) in which (x,y) and (p, q) are real-space and "Fourier-space pixel coordinates", respectively, of the same 2D image.
I have the image and its Fast Hartley Transform but can't really get beyond that. f(x,y) gives me the grayscale value of the point of course. p and q will identify a point in the FHT image but if I just use p and q as values, I'm getting an integer mask that has nothing to do with the image.
I don't think of Fourier space pixels as mapping 1 to 1 onto real space pixels so it doesn't seem to me that I'm looking for a corresponding Fourier space pixel for the real space pixel. That wouldn't work in k space anyway.
Any tips on this puzzle would be greatly appreciated.
|Register to reply|
|coordinates of phase space||Classical Physics||0|
|Hom. heat equation in cylindrical coordinates using Fourier & Laplace transforms||Differential Equations||1|
|Fourier transform in central and difference coordinates||Calculus||0|
|Simultaneous detection of the space coordinates of the ends of a moving rod?||Special & General Relativity||0|