Multi-camera global calibration for large-scale measurement based on plane mirror

Abstract

After analyzing and investigating the optical imaging geometry, a new multi-camera calibration method based on the planar mirror is proposed, combined with intermediary coordinates uniform method. By this method, the coordinate unification is achieved for two binocular vision pairs without overlapping view, which finally leads to the global unification of the whole multi-camera system. A multi-camera vision system is set up, including four CCDs, one mirror, and a plane template. Through global calibration by this new method, a displacement of 1100mm by a high-accuracy motor-driven moving platform is measured, with the average relative error 0.0205%, and the standard deviation 0.0740mm. The results show that the calibration method is simple and has high accuracy, and is suitable for precision measurement of large-scale measurands.