A Precise Calibration Method for the Underwater Binocular Camera With Unknown Refractive Index

Underwater camera calibration usually needs to solve an accurate refraction model to improve the calibration accuracy. In the refraction model, there are some key parameters: 1) the distance from the optical center to the refraction plane; 2) the thickness of the housing; 3) the normal vector of the refraction plane; and 4) the refractive index of the medium. In many cases, the refractive index is often assumed to be a known parameter. However, the refractive index of water will change due to environmental changes such as temperature, salinity, or depth. Inaccurate refractive index will seriously affect the measurement accuracy. Therefore, a precise calibration method for the underwater binocular camera with unknown refractive index is proposed. All the unknown parameters of the refraction model, including the refractive index of the medium, are solved by the nondominated sorting genetic algorithm (NSGA)-III. The parameter solving process is regarded as a multiobjective optimization problem, and the optimization objectives are determined by using the relative position and the actual position relationship of the checkerboard corners. Experiments show that the proposed method can greatly improve the 3-D measurement accuracy of binocular camera in unknown waters.