An automatic self-calibration approach for wide baseline stereo cameras using sea surface images

Abstract

The calibration process of a stereo rig is normally done by making two cameras view one special designed calibration pattern (such as checkerboard) at the same time to estimate the relative rotation between two cameras. Multiple image pairs from different perspectives are required for better estimation accuracy. Stereo rig with wide baseline is necessary when accurate depth estimation for distant object is desired. However, in order to make calibration pattern to be viewed from both left and right cameras, the wider the baseline the bigger the calibration pattern should be. Manpower even equipment such as crane is required to operate such a big checkerboard for the calibration purpose only. In contrast to the traditional stereo calibration method using calibration pattern, we propose a self-calibration approach that can easily estimate cameras' rotation matrices for stereo rig with very wide baseline using sea horizon and a point at infinite distance. Compared with conventional techniques which require expensive equipments or manpower, our approach only needs a pair of sea images, moreover, special calibration patterns and feature correspondences are not required. As a result, it is flexible and very easy to implement. Real world experiments demonstrate the feasibility of our approach.