malcircuit, Currently browsing Google Scholar to see whether it's possible to measure and model light pollution at a specific location in such a way to be able to calibrate it away.
Probably more trouble than it's worth (like most of my ideas), but I'm intrigued by the possibility.