A rough approximation for maximum telescope resolution at visual wavelengths (ignoring atmospheric turbulence) is:

a = 5.45/D where

a is the resolution in arc seconds,
and D is the telescope aperture in inches.

I.e., to successufully resolve a 1 arc-second separation binary (a separation of 1 Astronomical Unit (approx 150,000,000 km) at 1 parsec, or 3.26 ly) you would require a 5.45 inch (about 0.13843 M) aperture telescope--a modest amateur instrument. Think of this as a resolution of 1 AU at 1 parsec.

(Visual astronomy of the Deep Sky, Roger N. Clark, Cambridge University Press/Sky Publishing Corp).

Resolution is affected linearly by both the wavelength at which you are observing and the aperture of the telescope employed. You should minimize the first and maximize the second to increase resolution as much as possible. But if you double the distance of the observed object, the resolution drops (gets bigger) by a factor of two. At two parsecs (about 6.5 ly), you'd need an 11 inch scope to split that binary.

To achieve 1 km resolution (again, at visual wavelengths) at a distance of one parsec, you would have to have your two interferometer lenses 150 million times as far apart as the aperture of the scope in our example. This makes sense, if you want the object to appear that much sharper, you need a lens that much bigger .

So, converting to km, 1.5 x 10^8 x 1.3843 x 10^-4 = 2.0765 x 10^4 or 20,765 km apart.

Critter eye color may be too much to ask for, but continent-sized land masses don't seem unreasonable with forseeable technology.

BTW, would someone please check my figures?