05-19-2013 08:13 AM - edited 05-19-2013 08:19 AM
Hi everybody,
I have some problem with applying a convolution of the image and Smoothing Kernel in LabView. The idea behind this is, that convolution is a point-wise multiplication in the fourier domain. So firstly I want to apply convolution by standart IMAQ Convolute VI and then, to do basically the same, but manually make FFT of image and Kernel and then make their point-wise multiplication then apply inverse fourier transform and observe the result picture. But finally, I get two different results, but they should be the same, I can not get the idea why. I kindly ask to explain me, if smb. know the reason.
05-19-2013 10:27 AM
Sorry, found the solution. Just forgot about normalization of the core of convolution.
05-19-2013 12:02 PM - edited 05-19-2013 12:02 PM
Don't forget that the 2D convolution (frequency doman, size X) gives you about the same result.
05-19-2013 01:29 PM
Thanks in advance.