Calibration of Vision System Used on Unmanned Surface Vehicle

Vision system is the important equipment for Unmanned Surface Vehicle (USV) to find obstacles and monitor surrounding environment. There are popular calibration methods for vision system used indoor and outdoor on land. However, the USV has to be transported, hoisted and deployed, and calibrated parameters may be changed in the process. Consequently, it is desired to calibrate the vision system in running environment. Due to limited deck area, rock of the USV itself caused by wave, and wind, it is inconvenient to apply the typical calibration methods to USV. In this paper, we propose a method to calibrate the vision system automatically, using a calibration object made with a planar pattern attached to a cylinder. Experiments from simulated data and real images show that the calibration method applicable to USV and other unmanned system.