The study also revealed a surge in offshore energy development, with wind turbines outnumbering oil structures by the end of 2020. The researchers developed three deep-learning models to classify detected vessels, estimate their size, and identify different types of offshore infrastructure. The study monitored 15% of the world’s oceans where 75% of industrial activity takes place, paying attention to both vessel movements and the development of stationary offshore structures like oil rigs and wind turbines between 2017 and 2021.
Key takeaways:
- Researchers have used satellite imagery and AI to map human activity at sea, revealing that three-quarters of the world's industrial fishing vessels and up to 30 percent of transport and energy vessels are not publicly tracked.
- The study, led by Google-backed nonprofit Global Fishing Watch, used 2,000 terabytes of imagery from the European Space Agency's Sentinel-1 satellite constellation and deep-learning models to classify vessels and offshore infrastructure.
- The data revealed dense vessel traffic in areas previously untracked, particularly around South and Southeast Asia, and the northern and western coasts of Africa. It also showed a boom in offshore energy development, with wind turbines outnumbering oil structures by the end of 2020.
- The study's authors argue that this kind of detailed mapping is crucial for global conservation efforts and for managing the increasing use of the world's oceans. They also note that new machine learning tools are democratizing access to data and tracking technologies.