Regression for Ocean Wave Parameter Retrieval under Hurricane Condition Based on SVR
To address the challenges of limited and inaccurate wave observation data under hurricane conditions via satellite,Sentinel-1 IW mode SAR data were utilized to analyze the relationship of SAR features and wave parameters.After selecting 26 variables as inputs,a multi-parameter wave retrieval model was developed using support vector regression(SVR).The results of this model,including significant wave height,mean wave period,wind wave height,swell wave height,wind wave period,swell wave period,and mean wave direction,were compared with ERA5 data,NDBC buoy data,and outcomes from the traditional MPI method.The findings demonstrate that the SVR-based multi-parameter wave retrieval model significantly enhances the accuracy of wave parameter retrieval in hurricane conditions compared to theoretical methods.
万方数据
维普
