nanophotonics/nplab

reconstruct_tiled_image generates NaNs

Opened this issue · 2 comments

rwb27 commented

(from Will)
When the reconstruct_tiled_image function (from within camera with loc) calculates displacements it occasionaly gets NaN values which inherently causes problems down the line.
What do you think the best solution to this is?

Increasing the exposure can 'fix' this

That's probably just making the problem less likely, if this is just failed image correlations I think we should just to raise a custom error and state this?