Ask Your Question

Revision history [back]

Best practice approach to capturing whole paper document frame

I'd like to understand best practice options for digitally capturing the contents of a piece of paper, with precise positioning.

I have control over the format of the paper (typically will be A4) so can add markers etc but I need the contents to be captured at exact locations so that I can align the contents to a template for further processing. The approach will need to compensate for rotation, perspective correction etc as I want to target mobile devices.

I already have a solution which makes use of QR code detection to extract document corner points, followed by some simple rotation and OpenCV perspective warping. But it's a relatively slow process (circa 8-10 seconds) and has some slight alignment issues, so I'm looking for other potential approaches.