Image-based and risk-informed detection of Subsea Pipeline damage

Abstract As one of the most important assets in transportation oil and gas products, subsea pipelines are susceptible to various environmental hazards, such as mechanical damage corrosion, that can compromise their structural integrity cause catastrophic financial damage. Autonomous underwater systems (AUS) expected assist offshore operations personnel contribute pipeline inspection, maintenance, detection tasks. Despite promise increased safety, AUS technology needs mature, especially for image-based inspections with computer vision methods analyze incoming images detect potential through anomaly detection. Recent research addresses some significant challenges environments, including visibility, color, shape reconstruction. However, despite high quality images, lack training data reliable image analysis difficulty incorporating risk-based knowledge into existing approaches continue be obstacles. In this paper, we industry-provided propose a methodology address faced by popular methods. We focus on posed opportunities creating synthetic using risk insights. gather information anomalies, evaluate general approaches, generate compensate result from lacking data, evidence thereby increasing likelihood more inspection