Could they use the high-res orbiter photos, and the lower-but-still-really-good ground-based photos and use some sort of AI algo to enhance the ground-based ones?
The idea being that they have high-res reference photos that are a one-shot deal but can take regular earth-based ones auto-enhance them from now on.
It could then show changes over time in high res?
I'm showing my limitations here, obviously, but I know what I mean... it makes sense in my head :)
But for what purpose would an AI generative assisted image actually do for science? This is an issue I have with the gung-ho AI crowd that thinks AI should be used for anything and everything all the time.
Even if you trained the model against the most detailed images available. That data was a mere snapshot of the exact time it was taken which in some cases is decades old. If things are actually changing on these bodies, then using that stale data to update current images would actually be damaging to science as it would be attempting to make the current look like the old. No! We need to see what it looks like now for the comparisons.
Enhance! It can only go so far. Otherwise, you're just a low-rent Hollywood SFX team generating new worlds for whatever space opera you weren't hired to work on.
Sure, super-resolution exists in various forms and this is one. But this is an example of not actually adding any scientific information to the photo: extraplating data like that will never yield anything new or unexpected, so...
The idea being that they have high-res reference photos that are a one-shot deal but can take regular earth-based ones auto-enhance them from now on.
It could then show changes over time in high res?
I'm showing my limitations here, obviously, but I know what I mean... it makes sense in my head :)