This PhD project develops a multi-scale deep learning pipeline to quantify changes in tropical coral reef ecosystems. The workflow links fine-scale observations collected by citizens and autonomous surface vehicles with drone imagery and, ultimately, satellite data, so that local ecological information can be scaled up to very large areas across the Indian Ocean.

PLANCHA develops low-cost autonomous platforms for monitoring coral reef habitats in areas that are difficult to access or too expensive to survey regularly. By combining onboard image acquisition, automated image analysis, and environmental DNA, the project aims to map reef flats and outer slopes around La Réunion while documenting associated biodiversity.

SEATIZEN turns sea users into contributors to marine observation by equipping activities such as sailing, board sports, and diving with low-cost sensors and connected tools. The project builds a participatory network for collecting environmental and biodiversity data while also returning useful feedback to participants, including activity metrics and safety features. In its pilot phase in La Réunion, SEATIZEN supports lagoon mapping, image annotation, and automatic species recognition through deep learning.