[verum ago]

Also amazing that archeologists spend time poring over Google Earth images looking for things like this. How tedious. I wonder if AI has been applied to this.

There are efforts in that direction, not necessarily “AI” per se but image recognition algorithms. There’s also some websites that post imagery and let public review it and tag locations of interest for review.

One of the major issues at this point is that a lot of sites require elevation/contrast-enhancement to be recognizable, but the enhancement tends to introduce image artifacts with a lot of straight lines and sharp corners. Both the volunteers and computerized recognition programs then flag the artifacts as looking manmade.

It’s still quite promising though. The current issues are not fatal to the technique, they just need some improvement (and such enhancement techniques are very rapidly improving all the time). One of the biggest improvements will be to throw drone imagery into the mix- the higher resolution (vs satellite) and more easily orthorectified drone imagery will need less enhancement, and therefore suffer less issues related to enhancement.

[ProtectOurFreedom]

Multi-spectral imaging would be very useful, too. There are lots of things that can be seen in IR and UV light that the human eye cannot detect. I was reading here recently about the Pompeii scrolls turned to charcoal in the volcanic pyroclastic flow and how they can be carefully unrolled and read using invisible light wavelengths.

My dad was a project manager on the first multi-spectral earth orbiting satellites, originally named “Earth Resources Technology Satellite” (”ERTS”) and later renamed “Landsat.” Before that, he also worked on the first Nimbus meterological satellites launched in the 60s.

I remember Dad trying to explain “false color” images to me when I was in high school. As a young teen, I kept thinking “Huh, how can there be light we cannot see?”