The National Subsea Centre (NSC) and Robert Gordon University (RGU) in the UK have joined forces on a data project aimed at integrating artificial intelligence (AI) technology into subsea inspection operations.

Known as the Subsea AI Body of Knowledge (SAIBOK) project, it leverages AI and machine learning technologies to train algorithms for accurate detection and interpretation of anomalies in a collection of subsea inspection images.

Professor Eyad Elyan from RGU, who leads the SAIBOK project, explained, "The performance of AI models heavily relies on the quality and diversity of available data. Through the SAIBOK project, we have ingeniously utilized rich and diverse subsea data obtained from various industry partners for asset inspections, resulting in intelligent anomaly detection methods. This initiative has demonstrated the critical need for data sharing within the energy industry and presents a unique opportunity for companies to derive significant benefits from AI."

Supported by the Net Zero Technology Centre (NZTC), the project is a collaboration between NSC, RGU's School of Computing, and industry partners including TotalEnergies, BP, Chevron, Intel, and Xodus.

NSC suggests that the project could lead to the deployment of automated unmanned vehicles equipped with real-time anomaly detection and interpretation capabilities driven by the SAIBOK algorithms.

Blair O'Connor, Digital & Data Architecture Project Manager at NZTC, commented, "The SAIBOK project, driven by industry collaboration from the outset, sets an example for data sharing in the offshore energy sector. We have demonstrated its feasibility. The advancements in machine learning achieved through this project, made possible by access to a larger, multi-sourced dataset, have been truly inspiring. SAIBOK is delivering value for the project contributors, and the hope is that the broader AI development community will also benefit from the public release of the anonymized repository created for this project."