> Not that hard to imagine, OCR existed back then?
How do you train OCR on a character that doesn't exist though? Even these days, OCR often confuses the 52 Latin characters, so I wouldn't expect as good of results with older technology and some 20k CJK characters.
It does exist? It's part of Unicode!
It's listed in the Unicode character database, but I doubt that it was in (m)any fonts 20 years ago, it would have been in zero printed books [0], and nobody knew the meaning of the character, so I'd argue that it's more an artifact of the Unicode compilation process than a "real" character.
[0]: Aside from the "Overview of National Administrative Districts" book mentioned in the article
You don't need (m)any fonts, only need one font used by the book and you can get it the same way it was originally done - gluing some parts of other characters together. Or you could just draw it if the Unicode version is too dissimilar from the book version for a reliable OCR. (though if the character came from this single book, likely the Unicode reference is its replica?)
(Talking about Japanese here)
I don't know if this is how OCR would work here, but they're made of components called radicals. It's kind of like how letters make words, but in a two dimensional layout, and one letter was wrong and a nonexistent word was made.
For example 彁 is made of parts of 引 and 歌 (the first that popped into my mind).
IIRC there's only like 200-something radicals, though some can be hard to spot due to how they overlap or nest inside each other.
I find 鸚哥（いんこ） to be the easiest way to type the right hand side of 彁.