AI models & automation

Reef Support's AI models and automation service applies computer vision to imagery from satellite, drone, diver, and ROV sources to assess marine ecosystem health. The service covers coral reefs, seagrass meadows, and mangrove forests, producing quantitative cover metrics, health indicators, and change detection across survey epochs.

The system uses adaptive deep learning, processing reef assessments reported as 200 times faster than manual analysis.^[1] Input sources include satellite optical imagery for synoptic coastal context, subaquatic photo surveys for benthic classification, and drone or ROV imagery for structural assessment. Outputs include coral cover percentage and health indicators per site, mangrove and seagrass extent and health metrics, functional group assessments, and change detection across monitoring periods.

The service has been marketed under the names 'Open Coral AI' and 'Reef.io' in ESA BIC/Copernicus Masters period documentation,^[2] and as 'AI Models and Automation' on the current solutions page.^[3] Reef Support won the Copernicus Masters top prize in 2020.^[1] Applications include conservation planning, marine protected area management, restoration impact monitoring, and biodiversity credit and regulatory reporting (TNFD, CSRD, ESRS).