Shell Programming and Scripting

I'm trying to find a way to automate cleanup of OCR for a large number of scanned pages - due to limitations of the access mechanism where these are to end up, I need to create pdf files that include the background text for searching.

Going in I have Tif images too dirty to OCR and re-keyed text that matches page for page. I can see from reading here plenty of ways to turn the Tif files into pdf, what I can't find is a way to stick this text into the pdf file - I'm guessing this calls for some reverse-engineering of what ever mapping scheme pdf uses for the coordinates of words or characters. Does anyone know of a tool for getting access to this text - writing as well as reading. I'm looking at pdftk but so far all I can get is a dump of the "metadata" fields, but not the text with position mapping...

Are you looking to map each word in the manually generated page text to its corresponding position in the OCR image of the page?

There is an xml based metadata standard for this called METS-ALTO but what I'm trying to get at is the proprietary one that is inside of a pdf - the piece of the pdf file that is created when you run OCR within Adobe Acrobat.