|
Hi Jörn,
That is exactly what NDPluginROI is designed to do.
You create a plugin chain like this:
Camera -> NDPluginROI -> NDPluginFFT
You tell NDPluginROI the region of the camera you want to select. The FFT also behaves best if N is a power of 2, so 128x128 is better than 100x100.
Note that you can also use NDPluginOverlay to show the location of the ROI in the GUI that is displaying your camera image. That is demonstrated here:
Mark
Hello,
I have an areaDetector image which contains its relevant information in the center section of the image.
So to improve performance I would like to calculate the FFT of the image only for a subsection (configurable) of the NDArray.
Has anybody done something like that before?
I know, that FFT performs best if you have an array of size N * N, while most images are N * M with N != M.
So the best would be to create a sub array of lets say 100 * 100 pixels from the center of the image before doing the FFT.
Regards
Jörn
|