Originally Posted by
unterhausen
As a practical matter, I don't think it would be very good. Wheel2nut's picture is at some odd angle to the centerline of the frame. If it was square to the frame, then you would have a better chance. But you still are fighting lens distortion.
There are programs that try to make a 3d scan from pictures, I have no idea how successful they are. I would start there. But that's a real rabbit hole.
There is mathematics that can describe the correction from photo geometry to real object geometry so theoretically it should be possible to program it up. But I only know of the math, I don't know the math. and my experience in optics says the effect of lenses are also challenging to model. The math is called projective geometry, but I certainly don't know it well. And I don't have any sense of how it needs to be managed to create a model of known accuracy - without that my guess is you'll make something closer to what you really want or conceive just using 2D or a drafting table, though it may look well. I know I can draw the location of a hole much more reliably than I can drill a hole.
If you want the data for Wheel2nut's frame, get him/her to send it to you.