Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: None
-
Labels:None
-
Story Points:8
-
Epic Link:
-
Team:Alert Production
Description
Try out DFT space operations to understand:
- How to properly do convolution in DFT space in image and PSF kernel relation
- Padding, borders, translation of originals
- Numeric optimization of DFT, IDFT transformations by shifting significant values into the high-frequency region.
- Demonstrate a kernel inversion (deconvolution) operation, prototype how to do this in the decorrelation afterburner
- Ian Sullivan's hint: Check whether the above works for odd size FFT calls because odd size handling in DFT may be implementation dependent.
Write this in a jupyter notebook.
Based on numpy fft docs and fftfrwq, in the odd case the highest frequency is less than 1/2.