I personally did it based on my subjective visualization of light source (photometric measurement) but people do otherwise I guess - J Microsc. 2010 Sep 1;239(3):200-14. doi: 10.1111/j.1365-2818.2010.03366.x.

Photometric calibration for quantitative spectral microscopy under transmitted illumination.

Thigpen J1, Merchant FA, Shah SK

I would suggest you check the gain and exposure time of your camera to fine tune the image characterization. Say , if you increase gain then you can see ringing in the images, if you increase exposure then you might end up saturating pixels ... May be on set of illuminating constraints you can fix the parameters and use them as baseline !

Depends on what you are trying to do, and your setup. Maybe I misunderstand your question but for most systems you need to consider some or all of the following:

- dark current in the detector possibly by taking zero second exposures.

- dark images through the whole optical setup but without light source

- flat fields (uniform light source)

- and for photometric calibration you need to borrow/purchase a calibration lamp and possibly a calibration sphere.

The important thing is to take many measurements of each of these at many different exposure times. For example, taking many dozens or hundreds of dark images lets you measure the variance in the dark, allowing you to quantify error in your measurement. Taking at many exposure times can reveal non-linear response in your system.

Hope this helps a little, please contact me if you need more information. But possibly I misunderstand your question completely!