The spectral width of the sodium D2 transition is just 3 GHz- a VERY small fraction of the visible spectrum. The High pressure sodium lamps will broaden this width, but it will still be a very small fraction of the visible spectrum...

As for your idea about 'gating out' the the flicker of the lights- it is creative, but I dont think it would work in practice... The 60 Hz flicker would be roughly in phase near the power source, but two lights located far apart in a large city will be out of phase... also the lights will only be 'dark' near the null of the 60Hz cycle... (lifetime of the transition is , I believe, 16 ns, and would probably be shorter in a HPS lamp due to collisions I think that the 'dark time' would be quite short...)

Your idea IS similar to some proposed laser guide star systems... the idea is to use a precicely timed pulsed laser and gate your detector so you only observe the guide star when the laser pulse is in the sodium layer 90 km up... this allows you to avoid seeing the light scattered by the atmosphere as the pulse travels up through the lower atmosphere...