A Monocular Hybrid Visual Servoing Controller for Underwater Inspection on Work-class ROVs

This paper presents research and development for autonomous ROV manipulation systems with vision servoing in dynamic underwater conditions. Underwater inspection operations are performed by work-class ROVs equipped with robotic manipulators. The ROV generally travel at pre-planned paths extracting features from the environment, usually detected by a visual sensor. A monocular hybrid position based visual servoing (PBVS) system for underwater inspection on Work-class ROVs with a camera mounted on the manipulator and other on the top of the ROV is proposed. The effectiveness of the proposed system is verified through experiments carried out on real-world environmental tests consisting on using a work-class ROV to autonomously inspect a wall with 10 fiducial markers.