I would guess that you'd need 0.25mm resolution, which at 200km would require 250microarcseconds angular res, which doesn't even register on that graph at any wavelength or size of telescope lol
edit: this is a very 'spherical cow' analysis. i'm pretty sure that atmospheric effects will kill your ability to image stuff like this well before angular resolution will.
That chart is beautiful. While not really Tufte's dream of ‘a ton of measurements in little ink’, that chart shows me everything I'd want to know about this one relation.
Also, perhaps there's a need for a better naming system than one that produces ‘Extremely Large Telescope’.