Photometric Backgroud Extraction

dvolny1

Member
Is it possible to make relative Backgroud Extraction according to Star Photometry in RGB images?

The idea is:
- background is additive signal in image, it is added to the background and to the stars signal
- if you have enough stars in the image and you make star photometry on them, you can compute relative additive signal around many stars in image
- then you can use this to compute a background model which will be subtracted from image to correct the background

I don't know if the star photometry precision is good enough to compute a correct background model. But maybe somebody will know or try to find this out.
 
I think one problem with this is that stars are not always against the "background" (in fact, in many interesting images it is hard to find any real background). If a star is superimposed on a nebula or a galaxy, it doesn't really have a "background". I imagine it would be difficult to manage this problem automatically (it could be quite subtle in some cases).
 
In a way, this is precisely what NormalizeScaleGradient script is doing. However, it does this job based on a reference (making it a *relative* measure as the OP mentions but probably did not intend). So the "background" of stars is measured on superimposed signal- which is OK because you are comparing to a reference. So if you have "photometric" data (meaning, clear sky at low airmass) that is free from many issues- you can extract the relative backgrounds of other images and correct them. This spline-fitting gradient/background can even be outputted by NSG.
 
Yes, I thought NSG is working on a similar principle. Problem of NSG is that it is not really usable on objects that rise only let say15-20 degree above horizon. I was thinking that we can use scientific photometric data to correct our images even at low altitude where the light pollution (background) is worst.
 
Back
Top