Calibration of a 3D laser range finder and a camera based on stereoscopic checkerboard

In this paper, we propose a stereoscopic checkerboard based method for calibration of a 3D laser range finder (LRF) and a camera. Mixed pixels have great negative impact when extracting featured corners artificially, so we designed stereoscopic checkerboard to overcome this defect. By using depth division to separate the upper surface of checkers, we can successfully eliminate the impact of mixed pixels. As for the problem of point correspondences deviations which is caused by the LRF and the camera´s different resolution. We develop optimal combination method to find the best combination by removing point correspondences that have large deviation so as to improve calibration accuracy. At last, we utilize Levenberg-Marquardt nonlinear optimization algorithm to optimize results. Experimental results showed strong evidence that our method can achieve high calibration accuracy.