Calibration of stereo vision system used on autonomous surface vehicle

Abstract

Due to the special working environment, some problems should be taken into account for calibration of vision system used on autonomous surface vehicle (ASV). This paper discussed the relation between calibration parameters obtained from different distances. After a brief introduction to the imaging model of camera and the classical calibration methods, the relation between the distance from calibration rig to camera and calibration parameters is explored. By the least square algorithm, a functional relation is fitted. Based on equivalent transformation, a method to obtain the relatively accurate calibration parameters is proposed. Experiments show that the accuracy of calibration parameters is improved.