Multi-camera calibration based on freely moving one dimensional object

Since the one-dimensional calibration scheme has the advantages of simplicity, intuition, efficiency and high feasibility, we propose a novel method to calibrate internal and external parameters of multi-camera by using one-dimensional calibration wand. The method assumes that the principal point of each camera lies on the center of image plane. Then, using information of the fundamental matrix and constraints of distance between two feature points on the calibration wand, the internal and external parameters of a pair of cameras are simultaneously estimated by implementing a combination of linear solutions and nonlinear optimizations. After the pair-wise calibration, a scheme of global calibration of multi-camera system is proposed. Finally, simulation results show that this calibration method is quite robust and accurate.