AIR Centre showcases AI-based workflow to detect suspected marine plastic debris from space

The AIR Centre is taking part in the upcoming Commonwealth Clean Ocean Alliance (CCOA) webinar, scheduled for October 21 2025, 13:00 - 14:30 (BST) to showcase an AI-based approach to monitor floating plastic debris in the Honduras Gulf. Represented by André Valente, AIR Centre's Head of the Space Systems and Applications Program, this session will present a semi-automated workflow that combines Sentinel-2 satellite imagery, a deep learning U-Net classifier, and the dedicated POS2IDON data pipeline. This workflow generates insights into suspected plastic patches, including their spatial distribution and temporal dynamics in coastal waters. [caption id="attachment_21205" align="alignnone" width="1030"] Seasonal trends of suspected plastics derived from Copernicus Sentinel-2 at the Motagua River mouth coastal area (Gulf of Honduras) over a 6-year period (2019–2024)[/caption] Detection of plastic debris from space is of major importance for understanding and mitigating marine pollution. However, questions remain regarding the most effective detection methods, validation strategies, and integration into operational monitoring systems. In particular, there is strong interest in developing automated workflows that can provide complementary information to stakeholders. The study presents a methodology that can support stakeholders in assessing plastic loads, planning interventions, and evaluating mitigation measures, such as the OceanCleanUp Interceptor Barricade installed upstream of the Motagua River. The approach is adaptable to other regions, offering the potential to support long-term monitoring initiatives for marine plastic pollution. For more information about the event and registration, please visit the official webpage.