There are several levels to all this....
Quite helpful Frank.
Of course plate solving is just an initial step, not that the forerunner's of Astrophotographer stopped there.
Also, their goals were more of scientific discovery/evaluation than the objectives of a pretty picture that many amateurs pursue.
This leads to the realm of photogrammetry (making measurements from photos). And hence the creation of the term "astrometry."
Somewhere in the process of measurements for each image, there needs to be a correct identity found in order to catalog/classify, etc...
-----as this link with the associated photograph describes:
"The daily work was largely clerical: some women would reduce the photographs, taking into account things like atmospheric refraction, in order to render the image as clear and unadulterated as possible. Others would classify the stars through comparing the photographs to known catalogs. Others cataloged the photographs themselves, making careful notes of each image’s date of exposure and the region of the sky. The notes were then meticulously copied into tables, which included the star’s location in the sky and its magnitude."
One is just to have an image, or plate, and know the ID of each star because you know where it is aimed - and you have an atlas of stars in the area. If you are looking for variable stars or something - you don't need a "plate solve". You can just be handed a plate and knowing it's approximate ra/dec in the center - and plate scale - you can know what patch of sky it is - and then visually do the star matching to make the measurements and assign them to the correct star ID. I assume that is what is happening in the images above - and I don't think it's plate solving.
To know a star's location necessarily implies it was gathered from measurements in the image.
I have seen elsewhere the term "plate solve" in reference to particular objects in an image : i.e. "this object is identified as xyz as a result of the plate solve."
There is a distinction between a "blind plate solve" and one where most conditions might be known, but there is an unknown, yes?
Plate solving would be more for mapping - and involve knowing the exact scale, center, and orientation of the image - so that *unknown* stars could be measured and assigned ra/dec values. And back then I assume it would involve knowing the approx coordinates of the image - and the scale - and using known stars to figure out the plate orientation exactly.
Regardless, the process of astrometric solving was being done independent of the task (noticing variable stars, determining a stellar object's spectrum, etc.)...
Of course my posted MEME took liberty to illustrate my point about the term plate solving, regardless of whether the photo shows the ladies immediately doing such....
And of course some type of "plate solving" seems to have happened back then,
since the term "plate" was time specific......
Much more recently, with space missions and electronic imaging, there became a need to look at a pattern of stars and ask, "where the heck am I?" - because with a satellite in space the orientation isn't known a priori. And the key to that is a form of star pattern recognition - that looks at patterns of stars and maps them to known patterns of stars. And this was done early on by looking at triangle patterns of stars and creating a "hash" for them that was independent of scale, orientation, and mirror inversion.
It is interesting that you mention the use of star trackers on spacecraft and the process where they determine the craft's attitude. Funny enough, the term "plate solving" isn't commonly used at all in this arena. At most "solve" may be used.
I believe initially spacecraft used measurements from Sun detectors, measuring the Earth's magnetic field, and locking on and tracking available stars until they disappeared to gather data that was then sent to a ground station via telemetry and then the craft's position could be calculated with ground systems and then craft adjustments then could be relayed back to the craft.....
Of course as it became possible to add better computational electronics on-board, and with the innovation of better star detection capabilities in a small star tracker,
astrometric solving in order to determine attitude could be done onboard, quickly and adjustments could be made rapidly as well making things better in so many aspects!
More recently methods have become popular to use patterns of 4 stars in quads. And if done properly it allows a completely blind solve - where given an image anywhere in the sky - at any scale and orientation - you can identify the patterns of stars in the image - and then as a secondary result - provide a "plate solve" for the image.
That's what's happening when people center a telescope accurately based on "plate solves" - either using pattern recognition given hints of location and scale - or completely blind not knowing anything. But all that's really needed is the center RA/Dec value - which is only part of what a plate solve provides. And the overall process of centering based on that RA/Dec value requires much more than what's involved in a plate solve. It requires measuring the current RA/Dec, then making adjustments and trying again - etc.
Thanks for sharing an overview of the computational development of algorithms to identify a starfield!