But could images from several telescopes be combined through a Kalman filter or something to resolve higher details?
I am imagining some kind of internet enabled telescope that knows it's GPS location and orientation, and phones home it's imagery to a central server. If millions of people bought and used a product like that, is it theoretically possible to see the lunar rover?
TL;DR: for optical wavelengths, with typical image sensors that only detect amplitude and not phase of the electromagnetic wave, you need to do some really hard work to ensure optical coherency. For radio telescopes, it's a lot easier since you measure both amplitude and phase.
Another technique that's a lot easier to accomplish for amateurs is lucky imaging:
In this context, it's relevant to note that the VLT was built with interferometry in mind, and they are now getting it to work reliably. See for instance the page about the GRAVITY instrument: https://www.eso.org/public/teles-instr/paranal-observatory/v...
Here is a discussion thread with regards to the smallest lunar object visible - seems to be around 350 meter resolution for an earth bound telescope.
https://www.cloudynights.com/topic/712653-what-is-the-smalle...