The rapid development of artificial intelligence has created new opportunities for ocean remote sensing. Traditional methods still face limitations in representing complex sea states, integrating multi-source data, identifying extreme events, and retrieving ocean parameters with high accuracy. Recent advances in deep learning, foundation models, and data-driven approaches have provided new pathways for extracting ocean information and intelligently recognizing ocean phenomena. By learning multiscale, multimodal, and highly nonlinear features from massive remote sensing datasets, AI has significantly improved the efficiency and accuracy of ocean parameter retrieval, dynamical process identification, and marine phenomenon monitoring.

This presentation will introduce recent advances in AI-based retrieval and recognition of key ocean variables and phenomena, including sea surface temperature, sea surface winds, ocean waves, sea ice, mesoscale eddies, internal waves, and tropical cyclones. It will also discuss emerging trends toward intelligent, automated, and physics-integrated ocean remote sensing through representative applications such as multi-source satellite data fusion, super-resolution reconstruction, missing-data completion, target detection, and intelligent forecasting.