Underwater stereoscopic vision system calibration method

Abstract

The invention discloses an underwater stereoscopic vision system calibration method. The method comprises the following steps of: S101, in underwater environment, shooting images of a plurality of plane calibration plates by using a stereoscopic vision system formed by internal parameter-known cameras, extracting feature points of the image, and carrying out distortion correction on the feature points; S102, constructing a cost function according to a geometric relationship of refraction light paths, randomly selecting a camera in the stereoscopic vision system and calibrating refraction parameters of the camera; and S103, calibrating refraction parameters of the non-calibrated cameras in the stereoscopic vision system and external parameters of the stereoscopic vision system on the basisof the cost function according to the refraction parameters of the calibrated camera. By using the plane calibration plates, the calibration precision is improved; and on the basis of a multilayer refraction model, no system error is caused, the description precision is high, and the multilayer refraction model can be used for calibrating the refraction parameters of a single underwater camera orcalibrating the refraction parameters and external parameters of underwater stereoscopic vision systems and underwater multi-camera systems.

Description

technical field [0001] The invention relates to the technical field of underwater camera calibration, in particular to a calibration method for an underwater stereo vision system. Background technique [0002] In the underwater environment, the imaging light will pass through the three layers of refraction media: water, filter, and air. Existing methods often apply the traditional camera model to this refraction system, and use radial distortion coefficients to correct the deviation caused by refraction. However, the influence of refraction on the deviation of different pixels is different, and using the traditional camera model will produce obvious errors, especially in an environment with a high refractive index such as underwater. Traditional underwater camera calibration generally has the problems of complicated calibration methods, long calibration time, low calibration accuracy, and inability to perform underwater stereo vision calibration. Contents of the invention...

AI Technical Summary

Problems solved by technology

However, the influence of refraction on the deviation of different pixels is different, and using the traditional camera model will produce obvious errors, especially in an environment with a high refractive index such as underwater

Traditional underwater camera calibration generally has the problems of complex calibration methods, long calibration time, low calibration accuracy, and inability to perform underwater stereo vision calibration.

Images