ACKNOWLEDGMENT: REBECCA project is supported by the Chips Joint Undertaking and its members, including the top-up funding by National Authorities under grant agreement n° 101097224. Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the granting authority. Neither the European Union nor the granting authority can be held responsible for them.
REBECCA Use Case: Infrastructure Inspection with a Surface/Underwater Robot
Prepared by Almende
One of the use cases of the REBECCA project is a surface/underwater robot supporting both under- and above-water inspection of harbour infrastructure, developed by AquaSmart Engineering (ASXL) and Almende.
Automated inspection for infrastructure damage
The goal of this surface/underwater robot is to inspect harbour infrastructure like bridges, quay walls, sheet pile walls, and pillars efficiently and regularly to avoid maintenance backlogs. Traditional marine inspections often rely on manual visual checks by human divers, which can be time-consuming, costly, and hazardous. To automate and streamline this process, a novel AI crack detection model has been developed. This model analyzes visual feed during navigation to automatically identify structural flaws and generate actionable maintenance insights.
Testing across environments and initial results
The initial version has undergone rigorous testing across multiple stages: in a simulation, on a workbench, inside a controlled test container, and finally in the river Schie in Rotterdam. The testing phase yielded highly promising results and provided valuable real-world data under true field conditions.
AI model for crack detection
For defect detection and segmentation, Almende and AquaSmart Engineering trained a dedicated YOLO (You Only Look Once) detector model. Operating in maritime environments presents unique computer vision challenges: varying water clarity, fluctuating surface reflections, and organic growth on submerged concrete walls.
The developed YOLO model is specifically trained to filter out these environment-specific artifacts and isolate structural defects:
Defect detection and segmentation
Precision detection of critical structural issues, including cracks, corrosion, and spalling/missing concrete.
Driving smarter harbour maintenance
By combining AquaSmart's robotic platform with the developed AI models, routine inspections can be conducted more frequently and safely. Identifying early signs of concrete degradation allows port operators to schedule targeted preventative repairs long before structural safety is compromised, significantly extending the lifespan of critical harbour assets.