NavList:

Paul Hirose, you wrote:
"I tried the astrometry.net online "plate solver" with one of my own DSLR photos."

Fun, right? :)

The solver at astrometry.net is the "mothership". There are a large number of other astrometric "plate-solving" tools now available. I've occasionally considered creating one for my own use, but I haven't come up with a "compelling use case" [ so far ;) ]. Here's a partial list of some tools available: https://en.wikipedia.org/wiki/List_of_astrometric_solvers.

If you want to make your own, you may want to read an old but very clear article in the "Gleanings for ATMs" column in Sky & Telescope from Sep 1982 entitled "How to Reduce Plate Measurements" by Brian Marsden. There was also another article (maybe by Stuart Goldman) years later in the S&T computing column. That one included actual code, in line-numbered Basic probably, but simple to convert to any other language. Naturally there are much more modern resources with pages and pages of detailed code, but these are nice starting points.

There's certainly room for better tools in the astrometry category, and some almost certainly exist already in proprietary app suites. It's a basic "computer vision" pattern-matching problem. Older tools, like astrometry.net and any with roots in actual plate reduction, are poorly optimized for modern digital photography. It used to be understood that cameras produced images that behaved like gnomonic projections (I think it was Robin Stuart who mentioned this in a post within the past few months) or could be transformed with simple linear mappings. Given the power of digital photography to produce almost any projection, either directly on the camera or in post-processing, this constraint on projections probably has to be set aside or modified. The big problem, in my opinion, with astrometry.net is that it evolved rather "organically" into an all-purpose tool capable of analyzing anything from "snapshots" of wide areas of the night sky to arcminute-wide telescopic views. Any real-world application can be significantly optimized for narrower purposes.

You mentioned "photographic lunars". This should work... sometimes! When the Moon is more than about 30% illuminated, we get into a regime where the limb of the Moon and relatively faint stars near the limb are difficult to image simultaneously (then again, newer cameras are miraculous, and I wouldn't be surprised to learn that the latest iPhone or equivalent can blow past this limit). But when the Moon is a slim crescent, photos of the dark limb can easily show nearby faint stars, and an automatic astrometric solution should then be possible. Since that imagery would be taken at some known UT, the result, from at least two stars in the image, is a position fix in latitude and longitude with no horizon and no sextant required. Of course, this would require substantial magnification or exceptionally dense pixel arrays (or both). But it gets us close to that Holy Grail of automated celestial position-finding...

Now... if only we had some mini-moons orbiting much closer to the Earth. Heh. Yep, since 1957. The same process (as above) can be used to get extremely accurate position fixes from artificial satellites, potentially a thousand times more accurate than by actual "lunars", so long as we have decent ephemeris info for a few unambiguously identifiable satellites. I do hope that some military or intelligence organization somewhere on Earth has implemented this sort of visual satellite position-fixing system. It's been nearly twenty years since I first suggested it...

Frank Reed

PS: For anyone wondering... Plate solver? Why plates? The terminology dates back a few decades to the century-long era of photographic glass plates in observatories (see image below). The film emulsion was deposited on plates of glass, precisely because it created a stable surface from which exacting measurements could be taken later. The glass plates were typically 8-12 inches square. A developed plate looked like common clear glass with a little black dust scattered on it (the "dust motes" were images of stars --these were photographic "negatives"). Obviously these were delicate objects, easily broken or scratched. They required careful storage and curation. Many observatories still have vast libraries of glass plates. A large fraction have been digitized, but some have not, and who knows how many thousands of "transient objects" (novas, variable stars, lensing events, asteroids, comets) are waiting to be discovered in those old "plates"...

File: