Objectives of the service
Image credit: Sea surface vessel prototype, Italy Project: Zephyrus
The project aims to give MPAs (Marine Protected Areas), National Research Institutes, NGOs, and marine operators access to advanced technologies like artificial intelligence (AI), robotics, and space technologies that could help and accelerate marine conservation.
Specifically, the project will allow the autonomous data mining of coastal marine habitats through embedded AI on satNav localized underwater explorer-class drones and surface vessels, whose missions will be enhanced by cloud AI and Earth Observations (EO). The survey results and the derived insights will be accessible on digital twin dashboards for MPAs, which could share the marine habitat health status within their community.
At a local level, the proposed AI-empowered data mining solution will permit MPAs to effectively and cost-efficiently analyse key marine habitat factors over time and assess trends. These factors to monitor could be seagrass coverage and loss, benthic populations like corals or sponges, fishes -native or invasive ones-, plastic pollution – like ghost nets. Furthermore, it will allow National Research Institutes to quickly deploy best practices to the MPAs and coordinate actions to reach the aimed conservation goals at a national and international level.
Users and their needs
Through focused interviews with the various stakeholders of the marine protection (MPA, Governmental Organization, NGO, etc.), we have identified in particular the following main needs for the underwater drones and sea vessels:
- Customizable underwater drone with a different choice of setups and data mining capabilities
- Extreme drone usability for non-technical users
- High-resolution vision capabilities for benthic documentation
- Georeferenced data mining capabilities
- Underwater localization
Target areas: the Mediterranean Sea and all EU countries.
Service/ system concept
We propose a frequent monitoring and analysis service aimed at MPA leveraging semiautonomous satNAV-localized sea drones, custom AI and satEO companion analysis service. Overall, it will provide the MPA with digital twins of the natural reserves, containing frequently updated data, alerts, and actionable insights to cost-optimize conservation actions. It will be so composed of five sub-services:
Image credit: full-service architecture, Zephyrus Project
Space Added Value
The space added value to the service is many folds:
- The SatNav localization of the surface vessels via RTK positioning down to few cm levels will allow to precisely geo-localize and time stamp the data mined at sea;
- SatCom will allow communicating with the vessels at sea in areas not reachable by 3G connection;
- EO (optical data from Sentinel 2 or World View-2) will produce automatic change alerts over extensive areas for critical applications for marine conservation like monitoring seagrass coverage or anthropogenic pollution; given high-level objectives, like for example habitat degradation monitoring, the EO data will allow for the automatic definition of the best survey plans for cost-effective management of the MPAs.
We completed the kick-start activity achieving all the initial objectives of the project. We collected the user needs via interviews with end-users, from which we defined the service requirements. Given those, we successfully identified an overall service architecture, by splitting it into five sub-services ranging from AI model generation, to drones, vessels, and supporting app for MPA data-mining, to the companion EO AI analysis, to the end-point GIS and data dashboard. We then de-risked the most critical components of the service from AI and EO to the vessel and drone hardware, which we successfully tested in indoor and outdoor scenarios.
Image credit: Screenshot of control dashboard for a simulated data-mining mission, Zephyrus Project
Image credit: Synthetic Data for AI training needs, Zephyrus Project
We identified a grow business model and key partners to further develop the project in the next months.