A method, system, device, and medium for sea surface wind speed inversion based on time-delayed Doppler images and Swin Transformer

Abstract

A method, system, device, and medium for sea surface wind speed inversion based on time-delayed Doppler images and Swin Transformer are disclosed. The method involves constructing a time-delayed Doppler image DDM simulation model, outputting DDM simulation images, and building a DDM simulation image dataset. The DDM simulation image dataset is preprocessed. A Swin Transformer wind speed prediction regression model is constructed and trained using the preprocessed DDM simulation image dataset. The trained model is then used to predict and invert sea surface wind speed. A piecewise smooth activation function is designed to control the quality of the predicted wind speed values, thus completing the sea surface wind speed inversion prediction. This invention, by constructing a Swin Transformer wind speed prediction regression model and designing a piecewise smooth activation function, utilizes the self-attention mechanism and hierarchical structure of the Swin Transformer to achieve high-precision inversion over a wide wind speed range, exhibiting stronger robustness and physical interpretability.

Images