Since I started experimenting I have tried to improve calibration techniques for images taken at medium length exposure times 1-5minutes from my somewhat light polluted skies. The limitation on exposure time is the background ADU-counts getting to about 20-50% of saturation. With that much of background the actual signal (some nebula or galaxy) has a signal of just a few counts (<5%) above background. When I process the raw images to remove the background and utilize full brightness range for the object any slight deviation of the calibration frames (dark, flat, flat-dark) become obvious artifacts.
Here I tried flats and darks that were not taken at the same temperature than the raw images. The flat process also could have produced inhomogenities.
Looking at difference (distance)
images between flats taken under temperature drift
Over the cool down period of a night the CCD produced a sequence of flats. With progressing temperature the flats started to deviate from earlier ones more and more. Measure for deviation here was the distance function.
Looking at ratio images between
flats taken under temperature drift
Over the cool down period of a night the CCD produced a sequence of flats. With progressing temperature the flats started to deviate from earlier ones more and more. Measure for deviation here was the divide function. If adjacent flats would have been identical the ratio should have been 'flat' (i.e. constant 1). Apparently it is not.
Trying calibration with flats from
cloudy sky and first version of light box
I tried different methods to produce flats. Like a lighted disk or cloudy skies. Artifacts get smaller but still remain. over time again indicating a temperature drift. Cloudy sky seems to be better than light box.
Trying calibration with flats
from enhanced light box
I added more diffusers in between light box and telescope. I also tried to obtain flats and darks at similar temperatures than my raw images. Results are better now. However I believe there is no substitute for a temperature regulation of the chip temperature.
More attempts on calibration
with cloudy skies
After all a cloudy sky has given the best results for flats but some artifacts still remain. They seem to show up as a darkening of the calibrated image around the perimeter most noticably at the top lef and bottom sides.
Send me email at, firstname.lastname@example.org
Back to my astronomy home page