There is nothing random about that picture. It seems artificial and utile in some way.
"In some way" is not sufficient to establish that they are artificially constructed. Image-processing artifacts appear because the minute differences in contrast are artificially exaggerated.
They appear regular because they are magnified to the degree where the necessarily-regular spatial quantization becomes significant.
They appear regular in idiosynchratic ways because the processes of producing them lend themselves to certain characteristic and recognizable patterns of regularity.
What to a layman seems to be a jumble of pipes might to a petrochemical engineer be a fractional distillation apparatus. Unfortunately that means the layman lacks the knowledge required to distinguish a truly meaningless jumble of pipes from an actually useful apparatus. The layman can't say that something must be an oil refinery because it's a jumble of pipes, and to him all oil refineries look like jumbles of pipes. To call it an oil refinery requires having enough knowledge of the candidate identification to be able to identify actual functional components of it, such as identifying which components of the jumble perform which task, with a high degree of completeness.
It is unfortunate that DCT and error-diffusion artifacts resemble tracts of farmland when artificially enlarged and made to contrast greater. That doesn't mean they are farmland, especially when the farmland identification also requires evidence of a process actually undertaken by which the farmland pictures were made invisible such that the "enhancement" process revealed them, and the fantastic and improbable implications of farmland on Mars. It's more parsimonious to conclude that the artifacts are appropriate and natural to the photograph that was used as the color map for the 3D terrain plot, and that the "enhancement" process made them visible.
|