FFT of image on FPGA

[i.popa]

Hi,

I'm using a Basler camera with an FPGA as FrameGrabber. I'm obtaining images by using the example code 10-Tap-8-bit with DRMA.

I would like now to perform the FFT of the images I'm obtaining. I can already do it with IMAQ VIs, but I'd like to do it real time on the FPGA.

I tried to use the FFT Express VI for the FPGA, but I'm having some problems. The image comes as a 2D array, but the imputs for this VI must be single numbers. How I'm I suppossed to use this VI for FFT on images? Should I pass pixel by pixel to it, and late rearrange the pixels to get the new image??

Thanks for your help.

[Knoxberry]

You should pass the image in point by point, row by row. For example, if your image has a width of 1024, you would specify 1024 as the size in the FFT setup, pass each pixel in one at a time while the input is valid, and then wait for the output to be valid. After the output is valid, the FFT will pass out each frequency bin with the bin number specified by the output "data index". Once the output is no longer valid, you can begin the next row of pixels. If you want more reading material, see the following link:

http://www.ni.com/white-paper/7088/en/