Closed ollie-nye closed 4 years ago
ChArUco boards successfully used to orient cameras but tvecs don't appear to match up to what is physically measured so needs more investigation. Overall view is feasible though, if time allows
Closing as it's much simpler to just position cameras in known orientations
Using ChArUco boards, investigate the possibility of calibrating cameras from the position of the board.
Positional calibration would take place after fisheye undistortion has been applied, but at full resolution. Measuring the position of the board in space, combined with relative vectors from the board to the camera and rotation points, would massively simplify the margin of error for adding new cameras to the system.
This could also be extended to fixtures by identifying a hotspot produced by a fixture, and lining it up against the board on the floor. This would allow the system to know the elevation (from the tilt channel), azimuth (from the pan channel) and radius (from measured height combined with existing angles in simple trigonometry) that defines that fixtures' point in 3D space.