So, taking the standard diffraction limit calculation, I get a resolution of 15km using an orbital height of 250km and an aperture of 2.4m (as the KH-11 uses).

So, taking the standard diffraction limit calculation, I get a resolution of 15km using an orbital height of 250km and an aperture of 2.4m (as the KH-11 uses). That resolution is inversely proportional to the aperture. So you could get it down by taking larger apertures. The best way to do that would be a phased antenna array, but that is computationally very expensive. It would be very cool though.