Investigating the Ecological Impacts of Marine Debris on Fish Habitats Using High-Resolution Acoustic Cameras and AI-Driven Analysis

Marine debris, as a negative externality, is often regarded as a threat to aquatic ecosystems. However, its impact is not exclusively negative; some debris can function as artificial habitats under certain conditions, providing ecological benefits. Verifying this hypothesis requires long-term monitoring and systematic analysis. Given the low-visibility nature of most aquatic environments, traditional observational methods face significant limitations in uncovering the potential ecological impacts of marine debris. This study proposes an integrated monitoring framework that combines high-resolution acoustic imaging technology and machine learning algorithms to investigate the ecological effects of debris on fish behavior. Specifically, acoustic visual data on debris distribution and fish behavior are collected using a single acoustic camera. Machine learning algorithms are then employed to enhance the monitoring capability of the framework to distinguish between fish and debris targets. Finally, the optical flow technique was applied to visualize fish behavior trajectories around the debris. The approach was validated through water tank experiments and field tests, demonstrating its feasibility and robustness.