Samenvatting

Marine biodiversity is increasingly threatened by climate change, pollution, habitat degradation, and overexploitation, highlighting the need for accurate and efficient monitoring of seafloor habitats. Traditional methods of habitat classification and species identification are time-consuming, labour-intensive, and often subject to human error.This research project, conducted in collaboration with the Royal Netherlands Institute for Sea Research (NIOZ) as part of the European BioBoost+ initiative, focused on developing and evaluating an artificial intelligence model based on YOLO v11. The model was designed to automatically identify and classify seafloor substrate types and macroalgal species from underwater video footage collected in Dutch marine environments, particularly the Oosterschelde and Waddenzee.The research addressed the challenges posed by underwater conditions such as turbidity, sediment, and limited visibility, which reduce the effectiveness of existing AI models. The developed model enables automated analysis of habitat composition and macroalgal coverage with minimal manual intervention.The project aimed to assess the accuracy and reliability of AI-based habitat classification and algae coverage estimation. Ultimately, the technology could significantly reduce analysis time, improve the consistency of ecological data, and support large-scale biodiversity monitoring, marine conservation, and habitat restoration efforts.