Pushing subsea inspections toward full autonomy: inside the Velella project
At IVM Technologies, our mission is to make subsea inspection smarter, safer, and fully autonomous. Through the Velella project, we are developing solutions that combine autonomous navigation with vision-based photogrammetry inspection, moving a step closer to fully autonomous operations offshore.
Collecting the data: Floatgen mooring chain inspection
Autonomy starts with understanding the environment. Earlier this year, our team went onsite at the SEM-REV offshore test site, operated by Fondation Open-C, to survey Floatgen, a leading floating wind farm site designed and operated by BW Ideol, to collect mooring chain inspection data. This data is crucial for training neural networks capable of recognizing and interpreting subsea objects.
Using high-resolution imagery from inspections, our engineers created a visual recognition dataset, which forms the backbone of our algorithm. The dataset allows the system to:
- Detect specific parts of objects, such as mooring chain, connectors, and structural elements.
- Learn patterns of typical subsea structures to distinguish anomalies or defects.
- Enable vision-based control, where the ROV can make decisions based solely on what it “sees.”
This approach reduces reliance on pre-programmed paths or human intervention, making inspections adaptive and intelligent.
From data to algorithm: building autonomous vision
Once the dataset was collected, our team trained a neural network-based algorithm designed to:
- Recognize objects in complex subsea environments.
- Plan inspection paths automatically to capture all necessary views for 3D photogrammetry.
- Control camera orientation, ensuring that the imaging system collects optimal data from multiple angles.
This system essentially mimics what a human operator would do but faster, more consistently, and without fatigue.
Simulating autonomous photogrammetry
Before testing in the real world, we created a simulation environment based on our collected dataset. The simulation included:
- Hydro 300, stereo photogrammetry camera system, simulated with two integrated cameras on an ROV platform.
- 3D models of subsea structures, derived from the Floatgen dataset, serving as inspection targets.
In the simulation, the algorithm performed inspections fully autonomously:
- It navigated around the object along coherent paths.
- Tilted and positioned the cameras to capture all viewpoints necessary for high-quality 3D reconstruction.
- Generated a 3D model of the object from the captured imagery, essentially producing a simulated 3D model from an existing field 3D model, verifying the accuracy and efficiency of the system.
Looking ahead: real-world validation
The next step is deploying the algorithm offshore, in real-world conditions. This stage will test:
- Reliability: Can the system operate continuously without errors?
- Precision: Can it generate photogrammetry models of the same quality as manual inspections?
- Efficiency: Does autonomy reduce inspection time and operational costs?
Achieving these objectives will bring us closer to fully autonomous subsea inspection, transforming the way offshore wind farms, oil and gas assets, and underwater infrastructures are monitored.
Contributing to sustainable offshore development
The Velella project is part of a larger R&D effort to prepare floating wind farms for deployment under optimal economic, social, and environmental conditions. Funded by the French State under the France 2030 program and operated by ADEME, this project highlights the importance of innovation in sustainable energy infrastructure. It is also a way to innovate and collaborate with different French key players: BW Ideol, Centrale Nantes, Fondation OPEN-C, Ifremer and ENSTA.
By integrating advanced computer vision, neural networks, and autonomous robotics, IVM Technologies is shaping the future of subsea inspection, making it safer, faster, and more autonomous.
Implications for IVM Technologies’ future
The successful development of fully autonomous navigation and inspection algorithms marks a strategic milestone for IVM Technologies. Positioning and navigation have long been at the core of our R&D roadmap, underpinning the precision, safety, and efficiency of all our subsea inspection products. By integrating vision-based autonomy into our solutions, we are not only enhancing the capabilities of existing ROVs and photogrammetry systems but also laying the foundation for next-generation products that can operate with minimal human intervention. This evolution positions IVM as a leader in intelligent subsea inspection, able to deliver faster, more reliable, and cost-effective services to offshore operators, while continuously pushing the boundaries of what autonomous subsea robotics can achieve.
Reducing costs of subsea asset inspections
Contact us today to discuss your requirements, explore our solutions, and embark on a transformative journey in underwater inspection.