干涉合成孔径雷达多层次数据仿真

Multilevel data simulation of the interferometric Synthetic Aperture Radar

摘要:

为了对干涉合成孔径雷达(InSAR)系统研究提供数据源，提出了一种InSAR多层次数据仿真器(包括原始信号级(RSL)、单视复图像级(SIL)和干涉相位级(IPL))，该仿真器能同时适用于星载和机载合成孔径雷达(SAR)/InSAR数据仿真，并针对雷达后向散射系数在InSAR中的应用作了一定的改进。对每一层次InSAR数据仿真实现进行讨论：对于RSL仿真，通过利用信号生成的对称性，基于CPU的多线程处理和基于图形处理单元(GPU)的超线程处理等手段，大大提升原始回波生成速度；对于SIL仿真，提出了一种快速有效的InSAR复图像生成方法；对于IPL，提出了一种由几何去相关而引入的相位噪声模型。最后，实验结果证实了理论分析的正确性。

Abstract:

In order to provide the data source for the research of Interferometric Synthetic Aperture Radar(InSAR) system，a novel InSAR simulator was studied，which consisted of the Raw Signal Level (RSL)，the Single-look-complex Image Level(SIL)，and the Interferometric Phase Level(IPL). This simulator could be applied to the Synthetic Aperture Radar(SAR)/InSAR data simulation of both spaceborne and airborne. A new model of complex backscattering coefficients for the InSAR data simulation was presented and validated. Furthermore the implementation details of the multilevel simulator were discussed. For the RSL，a new model with higher computation efficiency to generate the radar echo was proposed，benefitting from the symmetry of the signal generation，the multi-core technique of CPU and many-core technique of Graphic Processing Units(GPU). For the image level，an efficient approach to create single-look image pair was presented. For the IPL，a reasonable noise model related to geometric decorrelation was also proposed. Some meaningful results by the novel simulator verified the theoretical analysis.