高分辨率星载SAR数据产品分级研究

有理函数模型RF在卫星图像领域的主要应用是进行对象空间和图像空间的转换。Radarsat-2元数据中使用三阶有理函数模型和80个有理多项式模型参数RPCs(Rational Polynomial Coefficients),为RF模型从光学图像领域的应用扩展到合成孔径雷达(SAR)图像领域提供了契机。本文对多种传感器、成像模式及不同处理级别SAR图像的RPC产品的制作和属性进行综述,主要研究高分辨率星载SAR卫星标准分级产品的制作流程。以中国广州作为实验区域,利用该地区不同级别的条带模式SM(Strip Map)Terra SAR-X和COSMO-Sky Med数据,对提出的分级和生产方案进行验证,结果表明,有理函数模型在SAR图像产品分级体系中具有可行性且效果良好。

Multiple auxiliary RPC products for the geometric processing of high-resolution spaceborne SAR datasets

The accepted Rational Function (RF) model application for satellite images is limited to transformations between object and image space. However, the use of the third-order RF model and of the 80 Rational Polynomial Coefficients (RPCs) in the Radarsat-2 metadata extends its application from optical images to Synthetic Aperture Radar (SAR) images. We provide a brief overview of the generation process and properties of multiple SAR RPC products, with special emphasis on the processing chain of basic SAR products in combination with orbit model approximation. Finally, we validate new types of RPC geometric performance experimentally. We consider the different basic products of TerraSAR-X, the COSMO-SkyMed of the Guangzhou area (in China) in stripmap mode as test data and demonstrated the effectiveness of the RF model for the geometric processing of SAR images.