@thomasfuchs But that's ignoring another aspect of the situation that might be used: we know (or can compute) the dynamic, geometric relationship *between the stars, the comet, and the camera*. We can compute where the comet *should be* in the image, relative to the stars. It's a third piece of information that could (maybe?) be used to "solve" the system and make an even more accurate image.
2/4