Benthic habitat mapping with autonomous underwater vehicles

Abstract

We provide an outline of an autonomous benthic habitat mapping algorithm. This algorithm enables real-time on-board classification of images gathered by an autonomous underwater vehicle (AUV), with the ability to classify aquatic vegetation at a resolution approaching the species level. The algorithm is generic with respect to both the classification task and the imaging equipment being used. For example, it may be used to detect objects rather than habitat types using hyperspectral rather than visible light cameras.