What configs to set Tesseract for text recognition?

Question

Python

I went through a lot of parameters, but I still couldn’t get all the text from the image.

pytesseract.image_to_string('time.jpg', lang='rus', config='--oem 3 --psm 13')

Answer 1

I won’t tell you about the configuration, but if you divide the area into 2 lines and apply filters to the bottom one, you can easily get the desired result.
Thresholding first, then invert:

And if you only need the bottom line, then you don’t even need to divide into regions.
On the web resource https://tesseract.projectnaptha.com/ it is successfully defined as "5 m."