Free and global pose calibration of a rotating laser monocular vision sensor for robotic 3D measurement system

摘要

A method based on a free-pose planar target is proposed to calibrate the position and orientation of a one-dimensional rotatable laser monocular vision sensor as well as its camera and laser plane parameters. For sensor parameters calibration, a new invariant in projective geometry, the invariable set of vector products’ orientations, is proposed to sort the control points in each image thus ensure the correct calculation of the control points’ correspondence between target coordinate system and image coordinate system regardless of the pose of the planar target. For sensor pose calibration, a kinematic model is established and a chain transform method is proposed, which enables the transform of images taken in various rotating camera coordinate systems into one same camera coordinate system. With an analysis on eigenvalues and eigenvectors of the transformation matrices, the position and the attitude angle of the sensor is represented in function of time. Verified by experiments, this calibration method has high-freedom operation, simple calibration procedure and ideal calibration accuracy.