This is a complex query because it depends on the PDF itself (and how it was created), whether it can be done or not. As a first attempt, I tried using the Adobe PDF PDF to HTML converter
http://www.adobe.com/products/acrobat/access_onlinetools.html
and then try to fix the HTML after the fact with something like tidy
http://tidy.sourceforge.net/
PDF , , - , , , JPG, - OCR PDF .
, PDF , , , , , , . , / .., JPG/GIF HTML-, , , , .