i also didn’t find any more dokumentation on the calibration process and found it really hard to get it working. thats why i am hoping that JCamCalib can help in the process…please anybody who tries it report back here.
now i remember: the OpenCV Library features some functions related to calibration which are also mentioned in their Reference Manual. i’ve got no idea about existing implementations, though.
an easier, yet undoubtly less accurate way of correcting for this kind of distortion would be: projecting (see /girlpower/slide projector) the original image onto the inside of a sphere with variable XY to Z scaling ratio, then putting a virtual camera inside the sphere to render a ‘corrected’ image.
Hmmm… The idea with the projection inside the sphere sounds good, but we have a flash attached to vvvv that will be controlled by data out of the contour tracker. So i think our computer will be have a lot of fun. We have to check it…
I think Undistort is THE solution for my problem. I attach a chess grid as .jpg for everyone who will need it in future. I´m working on this calibration wright now… He doesn´t recognizes all the spots for calibration. What can i do?
Finally i get it working. But it took me a while to check that i have to mount the camera so that it is locked and can´t move. Then i realised to put the paper really near to the camera so that the grid is screenwide… When you know that it´s no problem ;)