Description
For systems to become truly autonomous it is necessary that they be able to interact with complex real-world environments. In this article we investigate techniques and technologies to address the problem of the acquisition and representation of complex environments such as those found underwater. The underwater environment presents many challenges for robotic sensing including highly variable lighting and the presence of dynamic objects such as fish and suspended particulate matter. The dynamic six-degree-of-freedom nature of the environment presents further challenges due to unpredictable external forces such as current and surge. In order to address the complexities of the underwater environment we have developed a stereo vision-inertial sensing device that has been successfully deployed to reconstruct complex 3-D structures in both the aquatic and terrestrial domains. The sensor combines 3-D information, obtained using stereo vision, with 3DOF inertial data to construct 3-D models of the environment. Semiautomatic tools have been developed to aid in the conversion of these representations into semantically relevant primitives suitable for later processing. Reconstruction and segmentation of underwater structures obtained with the sensor are presented.