[SciPy-User] [SciPy-user] 2D Autocorrelation

[magnus_p]

I am trying to do a Autocorrelation summation of a bunch of images.
i.e iF( F(x)^2 ) is added for all images (F=fourier transform, iF =inverse
F).

The images contain 1 or 2 point-sources (i.e binary system), and shape
(64,64). 

what I tried first was:

def addAutocorr(self, cube):
    a = zeros((self.CUBESIZE,self.CUBESIZE))
    for frame in cube:
        a += ifft2(fft2(frame)*fft2(frame))
    return a

But it doesn't seems like it is correct, the source is not centered in the
middle in all frames and, not in the final frame.
Do I have to shift all the frames before adding them together?

Then I tried the scipy.ndimage.convolve / scipy.signal.fftconvolve with no
luck...

Anyone that have some advice?
-- 
View this message in context: http://www.nabble.com/2D-Autocorrelation-tp25530720p25530720.html
Sent from the Scipy-User mailing list archive at Nabble.com.