Well...as many will tell you prevention is better than a cure...Ideally acquire data when your target transits across the zenith will always be beneficial to minimise both sky glow and light pollution.
Light pollution typically harbours itself in yellow/green wavelengths (~510nm to ~590nm). In some situations such as sodium vapor streetlights, it can be higher (up to 610nm). This will result in a green cast over an image and reduce contrast. Reducing the green cast usually isn't too difficult, but to gain contrast is hard if the S/N is low - you'll only introduce noise in the process. An easy way of increasing contrast is imaging through a red or ha filter and using the data purely for luminance purposes. Neither filter are heavily impacted by light pollution or sky glow.
There are plenty of tools available to reduce and/or remove gradients caused by light pollution. If you are specifically after a method in PS, give this one a go.
http://starryvistas.net/Articles/Gra...tSubtract.aspx
It is quite effective. You can add to this routine by creating complex gradient maps should colour shifts occur from different directions. There will however come a time when only so much post processing can bring an image back to life. A lot can be said for collecting data under pristine dark skies.
