If there were any reasonable way to do it, the software would cost a lot. Your key here was _scanned_. This means Optical Character Recognition (OCR), a very imperfect science at the moment. You will need OCR software, and there's very little free OCR software out there, let alone any Perl bindings to it.

You'll need to convert the PDF to text with some OCR software FIRST. THEN running perl against it will be easy.