DaveStrickland
Add simple background subtraction to calibrated image processing pipeline
Add the capability to perform simple (i.e. scalar value) background subtraction of individual calibrated (or navigated) frames to remove image-to-image variance in the background level. This is of particular importance for assembling multi-pointing mosiacs.
- It should allow the use of a user-defined region over which statistics are computed.
- It should allow stars to be excluded from the image or region.
- It should (somehow, method TBD) allow the region to include nebular emission that is not to be subtracted away.
The region definition method needs to be determined. DS9 format region file, or other method?
This should be an optional part of the ApProcessor pipeline.
Is the goal of this to remove the background gradient of light pollution or rather eliminate artifacts like vignette and amp glow?
This is something I have in progress already, just haven't had the time to really get it done. The aim would be to statistically determine the parameters of, and then subtract, any source of low spatial frequency non-astronomical background in the images without subtracting away real but very faint nebular emission.
With multiple input images this gets more complicated, to use what overlaps what to determine a self-consistent set of image-to-image scalar brightness offsets. That has some established solutions, but its hard to understand.
In terms of iTelescope images the calibrated images supposedly have already been bias, dark, and flat field corrected. So in principal they shouldn't have amp glow or the vignetting-like effects of bad flat fielding. The problem is that in practice they don't seem to be done well enough to use the full detector field-of-view.