基于RADARSAT-2四极化SAR影像的海面风速反演

Ocean surface wind speed retrieve from RADARSAT-2 quad-polarized SAR images

选取成像于中国东海和南海海域的102景RADARSAT-2(RS-2)精细四极化SAR原始影像以及ERA-Interim风场,在各极化通道下利用相应风速反演模型开展风速反演的研究。CMOD5地球物理模式函数对VV极化SAR数据反演风速效果最好,可以获取海面高精度风速数据。对于HH和交叉极化SAR影像,利用RS-2 SAR影像和ERA-Interim资料对现有风速反演模型进行改进,并提出了用于HH极化风速反演的PR_CS极化比模型和用于交叉极化风速反演的CP_CS模型。结果显示:基于两种模型的SAR反演风速与ERA-Interim风速具有良好的一致性,利用PR_CS模型的SAR反演风速与参考风速的均方根误差为1.54 m/s,利用CP_CS模型的SAR反演风速与ERA-Interim风速的均方根误差分别为1.43 m/s(VH极化)和1.51 m/s(HV极化)。

In this paper,we selected 102 RS-2 SAR images and collocated ERA-Interim reanalysis wind information to assesses different sea surface wind speed retrieval model in China Seas.The result shows that CMOD5 geophysical model function is the optimal model for VV-polarized data,which can retrieve high precision wind speed from SAR image.For HH-and cross-polarized SAR images,we propose two wind speed retrieval models:PR_CS model for HH-polarized data,and CP_CS model for cross-polarized data.The results show that SAR-retrieved wind speed has a good agreement with ERA-Interim winds.The RMSE between wind speed from PR_CS model and reference data are 1.54 m/s.The RMSE between wind speed from CP_CS model and reference data are 1.54 m/s(VH-polarized)and 1.43 m/s(HV-polarized),respectively.