MARIOW – Maritime AI-Guided & Remote Operated Welding System

Project Overview and Motivation

The joint project MARIOW focuses on the development of an intelligent, semi-autonomous underwater welding robot for the maintenance of maritime infrastructure. The aim is to significantly improve the quality, efficiency, and safety of underwater welding operations through AI-supported image processing and modern robotics. Currently, such work is mainly carried out by divers, which is labor-intensive, costly, and technically limited. MARIOW addresses these challenges by establishing innovative solutions for automated maintenance and inspection.

Scientific and Technical Objectives

At Fraunhofer IGD, the main focus lies on the development of a robust underwater camera system with real-time image enhancement using AI algorithms (Autoencoder/UNet). The system enables precise capture and measurement of the welding process even under difficult visibility conditions such as turbidity, bubble formation, and extreme brightness contrasts. The integration of stereo camera setups and marker-based positioning allows millimeter-accurate determination of the welding head and weld seam, as well as the correlation of image and process data for quality control.

Technical Implementation and Innovations

Fraunhofer IGD developed and tested several camera and housing configurations that operate in real time on edge computing platforms (NVIDIA Jetson, Raspberry Pi). AI-based image enhancement was trained with domain-specific datasets and integrated into the overall system. In addition, algorithms for 3D distance measurement and interfaces for integration into a ROS2-based robotics and welding system were implemented.

Results and Benefits

The result is a functional prototype of an underwater image processing system that enables the visual planning, execution, and documentation of the welding process. The technologies developed can be transferred to other maritime inspection and maintenance scenarios (e.g., offshore wind energy, pipelines, port infrastructure) and provide both scientific and economic value for industry and research. The modular architecture facilitates further development and transfer to follow-up projects.

Supported by:

Project details

Project Duration

The project started in 2022 and was designed for a duration of three years. The final development and testing phase was completed in October 2025.

Project Partners

  • AMT GmbH
  • UNTERWASSERKRAUSE – MUTZECK GMBH
  • TH Köln, Institute for Materials Application – Welding Technology
  • TH Köln, Laboratory for Autonomous Systems
  • German Research Center for Artificial Intelligence (DFKI), Bremen
  • Fraunhofer IGD
  • Technical Business Consulting Dr. Karl-Heinz Kellner

Industry Sector

  • Maritime Industry

Further information

 

Digital Ocean Lab (DOL)

Find out about other technologies and applications at our maritime research and testing facility in the Baltic Sea, north of Rostock, the Digital Ocean Lab (DOL)

 

UXO Yard

Realistic Underwater Research and Calibration Facility

 

ODAS

The Platform for Intelligent Ocean Monitoring and Data Analysis