传统基于地球物理模型函数(geophysical model function,GMF)的全球导航卫星系统反射测量(global navigation satellite system reflectometry,GNSS-R)海面风速反演存在特征提取准确度低、模型复杂度高等问题。针对上述问题,提出了一种...传统基于地球物理模型函数(geophysical model function,GMF)的全球导航卫星系统反射测量(global navigation satellite system reflectometry,GNSS-R)海面风速反演存在特征提取准确度低、模型复杂度高等问题。针对上述问题,提出了一种基于卷积神经网络的GNSS-R海面风速反演方法。通过构建卷积模块自动提取时延-多普勒映射图像(delay-Doppler map,DDM)中的观测特征,特征融合模块将提取的特征与辅助特征关联,全连接模块将上述特征向量逐级映射到海面风速。以“捕风一号”卫星观测数据为例验证了上述方法的有效性,较传统GMF方法,风速反演精度在均方根误差(root mean square error,RMSE)和平均偏差(mean bias error,MBE)上分别降低了0.51 m/s和0.19 m/s,反演效果分别提升了21%和16%。试验结果表明:该方法能够有针对性地自动提取DDM特征,有效提高特征提取的精度,同时显著降低模型的复杂度。本研究为同类卫星各种地表参数反演提供了新思路。展开更多
全极化数据可以获取比单极化数据更多的目标信息,研究发现C波段交叉极化数据同样可用于海面风速反演。针对RADARSAT-2 Fine Quad模式具有全极化成像的特点,以我国东部海域为研究区,结合同极化数据和交叉极化数据反演海面风速模型,探究...全极化数据可以获取比单极化数据更多的目标信息,研究发现C波段交叉极化数据同样可用于海面风速反演。针对RADARSAT-2 Fine Quad模式具有全极化成像的特点,以我国东部海域为研究区,结合同极化数据和交叉极化数据反演海面风速模型,探究各种极化数据的最优风速反演方法。对于同极化数据采用地球物理模型函数(GMF)和极化率模型(PR)组合的方式进行海面风速反演,对交叉极化数据采用C波段交叉极化海面散射模型(C-2PO)进行海面风速反演,反演结果与ERA-Interim风场数据进行比较分析;此外,对Scan SAR模式交叉极化数据的后向散射系数与海面风速的关系进行探索分析。研究结果表明,RADARSAT-2 Fine Quad模式四种极化数据选用合适的模型均可反演出高精度的海面风速,其中VH和HV极化数据的反演结果基本相同,交叉极化数据反演风速效果好于同极化数据,同时,Scan SAR模式交叉极化数据的后向散射系数随海面风速的增大表现出一定的线性变化趋势。全极化模式数据在海面风速反演上表现出比单极化模式数据较为明显的优势,将成为未来海面风速反演的发展方向。展开更多
Aquarius is the second satellite mission to focus on the remote sensing of sea-surface salinity from space and it has mapped global sea-surface salinity for nearly 3 years since its launch in 2011. However,benefiting ...Aquarius is the second satellite mission to focus on the remote sensing of sea-surface salinity from space and it has mapped global sea-surface salinity for nearly 3 years since its launch in 2011. However,benefiting from the high atmospheric transparency and moderate sensitivity to wind speed of the L-band brightness temperature(TB),the Aquarius L-band radiometer can actually provide a new technique for the remote sensing of wind speed. In this article,the sea-surface wind speeds derived from TBs measured by Aquarius' L-band radiometer are presented,the algorithm for which is developed and validated using multisource wind speed data,including Wind Sat microwave radiometer and National Data Buoy Center buoy data,and the Hurricane Research Division of the Atlantic Oceanographic and Meteorological Laboratory wind field product. The error analysis indicates that the performance of retrieval algorithm is good. The RMSE of the Aquarius wind-speed algorithm is about 1 and 1.5 m/s for global oceans and areas of tropical hurricanes,respectively. Consequently,the applicability of using the Aquarius L-band radiometer as a near all-weather wind-speed measuring method is verified.展开更多
文摘传统基于地球物理模型函数(geophysical model function,GMF)的全球导航卫星系统反射测量(global navigation satellite system reflectometry,GNSS-R)海面风速反演存在特征提取准确度低、模型复杂度高等问题。针对上述问题,提出了一种基于卷积神经网络的GNSS-R海面风速反演方法。通过构建卷积模块自动提取时延-多普勒映射图像(delay-Doppler map,DDM)中的观测特征,特征融合模块将提取的特征与辅助特征关联,全连接模块将上述特征向量逐级映射到海面风速。以“捕风一号”卫星观测数据为例验证了上述方法的有效性,较传统GMF方法,风速反演精度在均方根误差(root mean square error,RMSE)和平均偏差(mean bias error,MBE)上分别降低了0.51 m/s和0.19 m/s,反演效果分别提升了21%和16%。试验结果表明:该方法能够有针对性地自动提取DDM特征,有效提高特征提取的精度,同时显著降低模型的复杂度。本研究为同类卫星各种地表参数反演提供了新思路。
文摘全极化数据可以获取比单极化数据更多的目标信息,研究发现C波段交叉极化数据同样可用于海面风速反演。针对RADARSAT-2 Fine Quad模式具有全极化成像的特点,以我国东部海域为研究区,结合同极化数据和交叉极化数据反演海面风速模型,探究各种极化数据的最优风速反演方法。对于同极化数据采用地球物理模型函数(GMF)和极化率模型(PR)组合的方式进行海面风速反演,对交叉极化数据采用C波段交叉极化海面散射模型(C-2PO)进行海面风速反演,反演结果与ERA-Interim风场数据进行比较分析;此外,对Scan SAR模式交叉极化数据的后向散射系数与海面风速的关系进行探索分析。研究结果表明,RADARSAT-2 Fine Quad模式四种极化数据选用合适的模型均可反演出高精度的海面风速,其中VH和HV极化数据的反演结果基本相同,交叉极化数据反演风速效果好于同极化数据,同时,Scan SAR模式交叉极化数据的后向散射系数随海面风速的增大表现出一定的线性变化趋势。全极化模式数据在海面风速反演上表现出比单极化模式数据较为明显的优势,将成为未来海面风速反演的发展方向。
基金Supported by the National High Technology Research and Development Program of China(863 Program)(No.2013AA09A505)the National Natural Science Foundation for Young Scientists of China(No.41306183)
文摘Aquarius is the second satellite mission to focus on the remote sensing of sea-surface salinity from space and it has mapped global sea-surface salinity for nearly 3 years since its launch in 2011. However,benefiting from the high atmospheric transparency and moderate sensitivity to wind speed of the L-band brightness temperature(TB),the Aquarius L-band radiometer can actually provide a new technique for the remote sensing of wind speed. In this article,the sea-surface wind speeds derived from TBs measured by Aquarius' L-band radiometer are presented,the algorithm for which is developed and validated using multisource wind speed data,including Wind Sat microwave radiometer and National Data Buoy Center buoy data,and the Hurricane Research Division of the Atlantic Oceanographic and Meteorological Laboratory wind field product. The error analysis indicates that the performance of retrieval algorithm is good. The RMSE of the Aquarius wind-speed algorithm is about 1 and 1.5 m/s for global oceans and areas of tropical hurricanes,respectively. Consequently,the applicability of using the Aquarius L-band radiometer as a near all-weather wind-speed measuring method is verified.
