If you studied compute science, then you must be pretty much aware of
the Moore's law. And you are certainly aware of predictions that
computer would never play chess well enough. Which means that any
speculation about robot's future limited abilities is groundless.

I see nothing special about your yellow soil example. With digital cams
ever increasing abilities, in 10 years you will have a remote picture
that is indistingusheable from what human is able to see on the spot.
Some obscure geologist sitting in the comfort of his desktop and
watching the transmission over the internet would notice something
interesting. Then you can fund a new mission *for a fraction of manned
mission cost*.

Even more likely, the amount of transmitted data in 10 years from now
would be so huge, that you have to employ a very sophisticated data
mining technique, in order to extract some useful information. No way a
trained Joe Doe geologist could be able to do that on the spot. Yes,
unglorified astronouts are just expensive technicians.