Research on Real-time Monitoring and Risk Assessment System for Ocean Heavy Metal Pollution Driven by Deep Learning

Aiming at the risk assessment problem of ocean heavy metal pollution, a deep learning-based risk assessment system was constructed. A three-layer evaluation index system including pollutant hazard degree, environmental sensitivity, and exposure risk was established, and a graph neural network risk prediction model was developed. The system achieved a prediction accuracy of 91.2% in the validation of 2,876 monitoring point data. A practical application case showed that the system successfully warned of three potential pollution events in a coastal industrial park, 48 hours in advance, providing important decision-making support for ocean environmental management.