You'll also need a sheet feeding scanner with a hopper on it, that'll scan both sides at the same time. Otherwise, it takes far too long. The software with the scanner will OCR it automagically and create a PDF. I scan at 400 dpi, which looks real sweet on a retina screen. There are a lot of settings to tweak on the scanner, some experimenting will get you the best results. Make sure you turn the double feed detection on.
Use some denatured alcohol to regularly clean the window and rollers, I also use a solder sucker to blow the paper dust out of it.
And lastly, you'll never get 100% of a book to go through cleanly. Just rescan the screw-ups, and assemble the result using pdftk (a marvelous tool). I also like to scan the covers separately in color and fold them in.
Times vary, but I can scan an average paperback in 5 minutes. Turning the sheets sideways makes it go much faster.
Great tip, thanks - presumably it also reduces misfeeds.
Would you recommend a particular scanner?
How does pdftk help with rescanned pages, does it magically know where to reinsert?
You have to explicitly tell pdftk what to do.
https://www.syncfusion.com/resources/techportal/ebooks/pdf
It explains use of iTextSharp, a C# library.