This paper describes a vision-based system for inspections of underwater st
ructures, e.g., pipelines, cables, etc., by an autonomous underwater vehicl
e (AUV). Usually, underwater inspections are performed by remote operated v
ehicles (ROVs) driven by human operators placed in a support vessel. Howeve
r, this task is often challenging, especially in conditions of poor visibil
ity or in presence of strong currents. The system proposed allows the AUV t
o accomplish the task in autonomy. Moreover, the use of a three-dimensional
(3-D) model of the environment and of an extended Kalman filter (EKF) allo
ws the guidance and the control of the vehicle in real time. Experiments do
ne on real underwater images have demonstrated the validity of the proposed
method and its efficiency in the case of critical and complex situations.