Black image after recalibration

[mcranium]

I installed the openflexure_microscope_software according to the instructions in the markdownfile of this repository (including the memory split option). When I run the openflexure_microscope --recalibrate command the program starts working and also saves a microscope_settings.npz file. However, I realised already in report of the openflexure_microscope --recalibrate command that the lens shading table gets filled with really high values (mostly 255). This then results in a black preview image and black exported images from the openflexure_microscope --no_stage command.
I am using the official V2 camera with the lens removed on a stereo microscope.
I am a scientist, not a programmer, I hope you can help me out here.
Best wishes!

[rwb27]

Hi @mcranium that is a bit strange - the really high values suggest that the image the sensor records is very dark, and it's trying to boost the outer values to make it uniform but running out of headroom. Filling the lens shading table with high values should make the images brighter rather than darker, so I think something funny is going on.

This repository is somewhat abandoned these days, as we run the openflexure microscope on a new software stack - https://openflexure.org/projects/microscope/install

That may be less than totally helpful though, as we tend to suggest it's easier just to install a pre-made SD card (from our website) so if you're trying to integrate it with something else it might be annoying.

If it's specifically the recalibration routine that's the problem, you can test that out by downloading and running the python file here:

https://gitlab.com/openflexure/openflexure-microscope-server/-/blob/master/openflexure_microscope/api/default_extensions/picamera_autocalibrate/recalibrate_utils.py

I'm afraid it doesn't have the loading/saving settings integrated, but if you watch the image, it should change from non-uniform to a uniform bright image. If the last thing you see before it finishes is a uniform bright gray image, it's worked properly. It might be worth trying that before debugging this set of scripts further.