As I've tried to indicate above, the type of pdf file that's generated by a scanner is (like a jpeg file) basically a photograph of your document. It 'knows' which bits are black, which bits are white and which bits are different colours, etc but it has absolutely no 'knowledge' of the actual text that's within it. (i.e. it doesn't 'see' letters such as "a b c"; it simply 'sees' black shapes on the page). So you CAN'T edit any text within the picture because there's no text information in it to start with.
The ONLY way you can get text from a scanner is to use the OCR function. That 'takes a picture' of your document and then runs that picture through a special program, which looks at all the black squiggles on the page and attempts to 'read' them as letters. The result is a basic text file, which will hopefully contain all the words and sentences from the document but which won't hold any other features of the document (such as pictures, logos, information about text fonts and sizes, etc).
For example, if you were to create a pdf scan from this photo
https://s0.geograph.org.uk/geophotos/05/30/79/5307988_8e1ea79c.jpg
you'd end up with a document which looked exactly the same but was in the pdf file format. Opening it in a pdf reader (or any other program capable of opening a pdf document) wouldn't enable you to edit the information on the road sign because it's a 'photo', not a text document.
You could scan the image using OCR software but that would just result in a file which read "M1 The NORTH Belfast 8". All other information in the picture would be lost.
It's just the same when, say, you copy that nasty letter that the tax man has sent you. If you choose pdf format, you'll end up with a 'photo' of the letter, where the text can't be edited because your computer doesn't 'see' any text; it just sees a picture.
If you use your scanner's OCR option, you'll get a text file containing all the threats from the tax man but nothing else from his letter. (e.g. the file you get won't include the layout of the letter, the HMRC logo at the top of it, the information that the amount owed is in red or that the font used in the letter head is different from that in the body text of the letter. It will contain just the bare text and absolutely nothing else).
It's simply NOT POSSIBLE to scan a document and end up with the non-text areas 'photographed' but with the text areas lifted out, run through OCR software, then dropped backed into place so that you've got editable text within the image. IT CAN'T BE DONE!!