Coral reefs are among the ecosystems with the highest biodiversity on our planet, but unfortunately also among those most vulnerable to climate change and other human impacts.
To understand how to reefs react to stresses by changes in conditions, and devise methods and policies to protect corals, it is imperative to implement large scale coral monitoring efforts.
Unfortunately, the water column significantly limits the resolution at which coral reefs can be monitored from satellite images. Therefore, precise characterization of reefs must be conducted field work. However existing conventional methods for coral reef monitoring are extremely resource intensive: the required manual effort by experts for analyzing the data limits monitoring at scale.
At ECEO, our research encompasses the development of the next generation of coral reef monitoring tools. Driven by machine learning, neural-network-based mapping and classification systems promise to increase the scalability of coral reef monitoring tools by orders of magnitude: they reduce the expert effort required, and can be used with affordable consumer-grade cameras instead of relying on expensive specialized equipment.
In particular, in a collaboration with the Transnational Red Sea Center (TRSC) hosted at EPFL’s Laboratory for Biological Geochemistry, and the TRSC’s partners in the Red Sea, our efforts focus on establishing a monitoring baseline across the entire longitudinal gradient of the Red Sea. To this end, our machine learning-based monitoring efforts have been deployed in the context of the [TRSC’s 2022 expedition to Djibouti](https://vimeo.com/764080310).
Partners