AUV localization using depth perception of underwater structures from a monocular camera

In this paper, we propose a method of AUV localization and mapping using visual measurement of underwater structures. Since the inertial navigation system (INS) of AUV suffers from drift, visual observation of fixed objects can enhance the localization performance. In a framework of pose graph optimization, depth map estimation of underwater structures and tracing of AUV trajectory are concurrently handled. Estimation of absolute scale is also resolved by exploiting incremental measurements of inertial sensors. The proposed method is validated by experiments performed in a structured basin environment.