With further playing I've sorted out a combination that seems to work (for me at least). I've found a free program - FreeOCR - that you can use along with a language module specific to Portuguese.
Download the program from http://www.freeocr.net/
Download the Portuguese language file from here (Brazilian Portuguese unfortunately)
This will download a packed/archive file which when you unpack it will give you a series of files beginning with "por." - unpack these files into the folder c:windowstessdata
Then, when you run the FreeOCR program, you should be able to go to the "OCR Language" option (top right of the screen) and select "por".
Once you have a scanned image of the page (using whatever scanning software came with your scanner; at ideally 300dpi), you can load this image file into FreeOCR, select an area to interpret, and then save the resulting editable text file either as a plain text file or export it into MS-Word. And of course, once you have the text, you can run it through Google for a rough translation.
You may have to tweak the file a bit, especially if the original page is creased or blurred, but it recognises all the accents, it even gets the little ° symbols that "officialese" Portuguese uses all over the place.