AI – Driven Marine Debris Detection for Ocean Conservation

Marine debris is a growing environmental crisis that impacts marine ecosystem's biodiversity and human livelihoods. Oceans are flooded with waste like plastics, due to improper disposal, illegal dumping, and limited waste management infrastructure. Traditional methods of detecting and managing marine debris, such as manual surveys or trawler-based cleanups, are labour-intensive, costly, and inefficient for large-scale application. With growing urgency, there is a need for scalable, automated solutions that can monitor vast oceanic areas and provide insights for debris mitigation. To overcome these challenges, we have introduced an advanced system utilizing the YOLOv8 deep learning model which is trained to detect debris under challenging underwater conditions, including low visibility and varying light levels. By leveraging cutting-edge object detection techniques, the system identifies and classifies various debris types from underwater images with high accuracy. The project begins with data collection, representing an extensive dataset of diverse debris categories. Evaluation metrics like precision, recall, and F1 score demonstrate the model's efficiency in identifying debris while minimizing false positives. This model can decrease the intensive labour process and increase the efficiency for large scale applications. This research advances marine debris management through deep learning, promoting sustainable ocean practices and contributing to global environmental health. Future work involves extending the system's capabilities to detect microplastics and integrating it with real-time response systems for effective cleanup operations.