1. This is not a proper approach to seeing each PDEElement. You look only inside the top level PDEContainer, but the PDEContainer can hold PDEContainers which can themselves hold PDEContainers. Recursion is mandatory if you want to see all elements, but I don't know why you still use PDE at all.
2. You get a PDWordFinder and do not use it. There is no point in this code.
3. Expect PDF files to contain much text and many images. There is no special type of PDF which is "OCR'd". It is not different from other PDFs containing text and sometimes images.
4. It is possible for you to quickly say that a file contains text or no text. You can make decisions on this information. There is no easy way for you to know if text is "searchable", only whether it exists.