ENH more automation of flatfield correction

[ambarb]

see discussion in #77 for minimum_ADU suggestion. For FCCD, beamline staff recoomend collecting data wiht average 1000 ADU per pixel for a robust flatfield data set.

[ambarb]

we are getting a new ccd so probably not worth moving forward on this (or at least waiting to see if it can be done "universally")

if we set a threshold, then we probably need some extra arguments if the threshold cannot be met (i.e., we tray to have the mean ADU per pixel be >1000. Not always possible, and if the distribution of photon events overlaps with the dark noise, then we need to use a cutoff based on the distribution