Comment I worked on the project back in '68 (Score 3, Informative) 29
Back in ny early days as a lab techie I was running the optical processor that did the image-making post-processing for what I believe was the first "flyby" / "rotating target" synthetic aperture radar. (No significant intellectual contributions: I was running the machinery, rather than contributing to its design. Adjusting lenses, exposing and developing film, etc.)
Back in those days the computers weren't up to the amount of crunch needed. (This technique is essentially a two-dimensional fourrier transform with tweaks.) So we used laser light and lenses for the fourier transform, and photographic film for the input modulation and output capture. The original data was captured using a one-dimensional CRT with a solid row of fiber-optic light-pipes built into the faceplate. This was in actual contact with the recording film, transferrig the light from the phosphor inside the CRT without geometric distortions from lenses and such. The film was about four inches wide, and the servo capstain that advanced it was a critical component for accurate signal processing, as was the circuitry that linearized the sweep of the beam. The input plane of the optic processor held the film in a xylene solution between two optical flats, to eliminate phase distortion from roughness of the film's surface.
The nice thing about synthetic aperture radar is that the resolution is related to the radar frequencies and the relative motion of the antenna(s) and target, and is not dependent on the beam width of the antenna. (Well, wider beam width means you illuminate the target from a larger virtual antenna, sharpening the image.) Except for deltas, distance doesn't matter, either. You get the same resolution at tens of feet or interplanetary distances. Distance only comes into the pricture in terms of keeping the oscillators from drifting during the transit time of the beam, so you don't introduce varyimg phase errors when down-converting successive returned chirps.