If anyone disagrees with my points, please point it out, because I think we're stepping into dangerous territory in which it is difficult to distinguish between what is a "puzzle" and what is not. Also, I think I'm being quite harsh to this genre of puzzles, so I'd like to see where precisely any disagreements may be. Anyways, here are my own two cents:
Suppose we manage to disallow the use of a reverse image search. Then the "puzzle" is using clues in the image to deduce the location, as it is intended. From this point of view, this may not be a strict "trivia" question.
However, I think it is still problematic as it may be extremely difficult to deduce anything unless some trivia is applied, making it a trivia question anyway.
Allow me to examine all the existing puzzles to make my case. Take the first one, which turned out to be the Pearl Harbor Memorial:
Where is it? - The Google Earth Challenge Ep. 1
There were really no clues involved, unfortunately. Either you recognized the memorial or you didn't.
Now let's consider the second:
Where is it? - The Google Earth Challenge Ep. 2
This was slightly better, as you had to recognize that this was some sort of water attraction. Though, I'd venture that this is closer to a blind hopeful guess than a deduction, and to be truly certain that it was some sort of water attraction, you had to have seen something similar. IMO, this leads toward "trivia" than "puzzle". (What, I believe, supports my assertion here is the claim that "this was obvious" in Belhenix's answer) Even after this "deduction", what are you supposed to do? Look through random fountains or lakes that do cool stuff until you magically get it? How was I supposed to deduce that this was outside a palace, or in Macau?
Now let's look at Panama Canal:
Where is it? - The Google Earth Challenge Ep. 3
The answerer literally just guessed that it was Panama Canal, and then looked through Google Earth to confirm the guess. How is this a puzzle?
As for Athin's septuple puzzle (Seven Places at Once - Another Google Earth Challenge?), there really was nothing remarkable about this locations that could allow for any nice "deductions". Really, the only remarkable thing was the names of the locations.
As for Rand al'Thor's puzzle (Not Seven This Time - another Google Earth challenge), I think it was a good attempt at a location that could be found using some legitimate research. I'm unsure if this balanced out the amount of guesswork that may be needed.
Instead of a Google Earth puzzle where you have to deduce a location based on a satellite image, I suggest a puzzle that takes some screenshots (or just one, depending on the puzzle) from the Google Street View (or even combine them with the satellites) that have a significant amount of information within that one can use to deduce the location. Sort of like GeoGuessr, except the location is fixed and handpicked to be interesting.