Are there normal libraries for recognizing simple characters?

Question

Text recognising

Hello!
I am currently doing license plate recognition for my project. Initially, I thought that the most difficult thing would be to find an area with a number and segment it into characters, and later I planned to use a ready-made library (tesseract, gocr) for character recognition. But unfortunately the results turned out to be deplorable, they confuse 8 and B, U and L, although the characters look very clear and 0 with a serif in the corner does not recognize at all, although this font on the reference image recognizes all characters perfectly. Is there really no library at present that would cope with such seemingly simple characters, or do you still have to do it yourself? There are of course paid services, but I would like something simpler and more autonomous.

Character examples:

Answer 1

SchA everyone is engaged in neurons. But it needs to be trained - to generate a bunch of images of letters with the necessary fonts with given distortions, the angle of rotation, perspective angles, and other optical distortions are acceptable. In principle, this is not long and there are plenty of examples on the net.