AI-Driven Marine Citizen Science: SECOSTA’s Blueprint for SeabedExploration

The Balearic Islands are home to a rich diversity of seabeds, yet their distribution and evolution remain challenging to map and study, which has become a priority in the EU-marine strategy directive. The SECOSTA project contributes to bridge this gap by integrating advanced machine learning with citizen participation, driving both scientific progress and community engagement.

Central to SECOSTA’s success is its hands-on, co-creation approach, where high school students actively design, build, and deploy low-cost, innovative tools for marine research, including beach profilers, bathymetric probes, tide gauges, and bathythermographs. A standout example is the Arduino-based seabed exploration platform, collaboratively constructed by students under the supervision of the SECOSTA team. This device integrates a GPS chip, datalogger, and submersible camera mounted on a floating platform, enabling efficient collection of high-resolution, geo-referenced seabed imagery in shallow coastal waters. Designed for ease of use, the platform can be towed by a small craft, such as a kayak or paddleboard, or by a swimmer, allowing students to gather invaluable data on underwater habitats.

The project focuses on classifying and characterizing critical marine ecosystems, such as Posidonia oceanica, alongside benthic species, sediment patterns, and marine debris. Students label collected images using Roboflow to build a robust dataset, which is then used to train a convolutional neural network inspired by U-Net, a leading architecture for image segmentation. By engaging in every step—from designing the tool to enrichening the dataset used to train the AI—students gain a deep understanding of both scientific and technological processes, while developing a sense of ownership over the outcomes. 

Since its launch in 2018, SECOSTA has engaged over 7,500 students from 33 educational institutions, generating actionable insights for coastal management and fostering longterm community capacity. These achievements have been made possible through close collaboration between researchers, students, and local government, highlighting the importance of transdisciplinary partnerships in addressing complex environmental challenges. By blending participatory methods with cutting-edge AI applications, the project exemplifies how co-creation can empower communities to take an active role in tackling issues like climate change and biodiversity loss.

This presentation will explore SECOSTA’s co-creation methodologies, the technical specifications of its seabed exploration platform, and the lessons learned from integrating students into environmental monitoring and AI-driven marine research. SECOSTA exemplifies the transformative power of citizen science, where education, technology, and sustainability converge to inspire the next generation of scientists and stewards of the natural world.