Robust Camera Calibration for an Autonomous Underwater Vehicle

Abstract

At the Australian National University we are developing an autonomous underwater vehicle for underwater exploration and inspection. One of our aims is to track the relative position of underwater targets. This has required the development of a camera calibration system that can overcome the difficulties of underwater· vision to provide accurate camera parameters. Conventional camera calibration systems detect and then identify points in an image of a known 3-D calibration pattern. Point identification algorithms typically require the full set of calibration points to be detected to register the target, but this requirement.is seld~m satisfied in underwater images. We descrIbe a pOInt identification algorithm which does not rely on complete point detection, based upon the indexing of planar invariants calculated from points on a 3-D calibration pattern. Underwater experiments have shown our new method improves the likelihood of successful calibration by up to 80%, and that our calibration system calculates camera parameters enabling range estimation of targets up to 3 metres away with 95% accuracy.