Spatiotemporal Prediction of Real Ocean Physical Environmental Factors and Dynamic Multi-Scale Sea Surface Geometry Modeling Based on AI

The prediction of ocean physical environmental factors has important applications in radar detection, target recognition, remote sensing, telemetry, communication, and navigation. However, due to the extensive coverage and multiple dimensions of the ocean environment, traditional methods find it challenging to achieve accurate and real-time spatiotemporal predictions of ocean physical environmental factors. This paper utilizes models based on Deep Neural Networks (DNN) and Long Short-Term Memory networks (LSTM) to achieve spatiotemporal prediction of ocean physical environmental factors in different sea areas of China. By collecting multi-source and multi-dimensional ocean spatiotemporal physical field data, we constructed prediction models for ocean environmental physical fields using LSTM and DNN. The proposed models demonstrate faster speeds, better accuracy, and greater stability compared to traditional methods in predicting various ocean physical environmental factors such as wind speed, wind direction, wave height, wave direction, wave period, and temperature. Utilizing a real ocean environmental factor database combined with high-performance GPU technology, we established dynamic multi-scale sea surface geometry models for different sea areas based on different radar electromagnetic parameters. This provides theoretical foundation and technical support for ocean environment situational awareness, radar target detection and clutter suppression, high-resolution satellite remote sensing, and over-the-horizon communication.