fSpy
fSpy copied to clipboard
Multiple axis references / Least Squares approach
Sometimes it's not easy to find usable lines for an axis in the image. Maybe I can find a couple of parallel lines, but they are short, and the measurement error is large as a consequence.
Would it be possible to give the user the option to add more than two axis references (three, ten, fifty...) and to solve the camera matrix using a least squares optimization? The results wouldn't be "analitically correct", but this could help reduce measurement errors (especially because the image cannot be zoomed).
I can volunteer some coding help if it's needed.