Recently, Coastal Engineering published an online article entitled “Satellite wave 2D spectrum partition based on the PI-vit-GAN (physically-informed ViT-GAN) method”. The corresponding author of this study is Professor Tao Aifeng of CEcode.

This study proposes a Physically Informed ViT-GAN (PI-ViT-GAN) automatic partitioning method, based on CFOSAT satellite wave spectrum data. Specifically, the model consists of a generator and discriminator. The generator utilizes a contrastive learning strategy as pretraining and through the self-attention mechanism of the ViT model, it focuses on key parts of the spectrum to extract wave group features and wave element parameters.  Partitioning-head joint training realizes the output of wave group partition element indices.  Subsequently, the discriminator uses the wave group features and a parametric model for spectrum reconstruction and computes the error with the original observed spectrum to evaluate the partition and reconstruction effects.  Additionally, this model incorporates two physically corrected functions, wave system classification loss and merging loss, based on the wave age criterion, thereby guiding the training process, and enhancing model efficiency. The results indicate that the reconstructed theoretical spectrum, obtained through the utilization of this method, aligns well with the original sea wave spectrum, demonstrating a precision superior to the spectral partitioning product of CFOSAT’s own SWIM. This work not only demonstrates the potential for deep learning techniques to be applied in the field of Marine and coastal engineering, but also provides a new and scalable approach to big data analysis. This method can help researchers better understand and predict the complex physical processes in the Marine environment, thus providing a scientific basis for the design and implementation of coastal protection measures.

Article link: https://www.sciencedirect.com/science/article/pii/S0378383924000668