How to OCR?

[joseph22]

With Kubuntu 22.04, How to do OCR?
OCR = Optical Character Recognition

Testing done with Skanlite Version 21.12.3

Skanlite is outputting Ok to x.pdf

1.
Can Skanlite make x.pdf OCR?
text superimposed on an image

thus x.pdf is searchable for Text
meaning Ctrl-F = Find text inside file

and

thus x.pdf is indexable
meaning words inside x.pdf can be
indexed by Recoll version 1.32.7
Recoll is a desktop full-text search tool.



2.
If Skanlite does not have OCR
then can OCR be added to Skanlite?



3.
Suggestions
How to do OCR with Kubuntu 22.04 ?


--

[claydoh]

With Skanlite? No.
With Xsane, yes but you need to install an OCR util such as Gocr or Tessaract, and probably a lot of fiddling.

To have a PDF file with searchable text you will have to make sure that the PDF is not created from an image of the desired document, but created from the actual captured text.
There probably are Linux utils to help simplify this process, but they will often be command line or old.
Your phone probably will have superior tools and capabilities for this task, as an option to consider as well

[jlittle]

I use pdftotext from the package poppler-utils to great effect to extract information from some utility bills. The order the text appears is a mess, very different to the document, but there is just sufficient context to allow a script to identify what I need to load it into a spreadsheet. But, as claydoh, warns, some pdfs have just images.

[GreyGeek]

The latest, Tesseract 4, is in the repository. Or, you can install it from github.
Tesseract defaults to English, but if that is not your native tongue then be sure to install the language option for your needs.
Yagf is a GUI interface to Tesseract.

Cuneiform, also in the repository, is another tool set to use for OCR. Yagf can be used as a GUI for it as well, among others. Normcap, another Tesseract GUI, comes as an AppImage. Download it, set its execute permission and click on it to run it. Merely deleting an AppImage totally removes it from your system.

There are online OCR tools as well. OCR.net is one such tool.