A Flying Tool for Sensing Vessel Structure Defects Using Image Contrast-Based Saliency

Abstract

Vessels have to undergo regular visual inspections in order to detect the typical defective situations affecting metallic structures, such as cracks and corrosion. These inspections are nowadays performed manually by ship surveyors at a great cost. Assisting them during the inspection process by means of, e.g., a fleet of robots capable of defect detection would, without doubt, decrease the inspection cost. In this paper, a robotic sensor for visual defect detection on vessel structures is presented. Its hardware comprises a micro-aerial platform, a suite of navigation sensors, and vision cameras, to provide a close-up view of the vessel surface under inspection. The software architecture has been designed around the supervised autonomy paradigm, so that the user is introduced in the position control loop, while the platform is in charge of providing functionalities, such as collision prevention and other self-preservation issues. The inspection tool is completed with a vision-based defect detector that labels those areas that are suspicious of being affected by corrosion or cracks. Both hardware and software architectures have been assessed for the vessel inspection task with good results.