GPU-based SAR raw data simulator considering scattering centers' RCS variation

Cheng-qi ZHANG,Hui SHENG,Qian-ni DENG
DOI: https://doi.org/10.13274/j.cnki.hdzj.2017.04.031
2017-01-01
Abstract:A GPU-based synthetic aperture radar(SAR) raw data simulator is proposed in this paper.To obtain high-precision SAR raw data,the radar cross section (RCS) variation of SAR scattering centers in the illuminated scene,which is obvious especially when the angle of incidence varies a lot during the synthetic aperture time,is considered.Under high-frequency prediction,the backscattering coefficient of scattering center is formulated as a function of the incident angle and instantaneous slant range.Thus,given the explicit digital elevation model (DEM) information of illuminated terrain and the range of incident angle,the backscattering coefficient of scattering centers in different azimuth time can be obtained.Then,the hybrid GPU-based and CPU-based SAR raw data simulator is introduced.In this system,the introduction of matrix symmetry greatly reduces the computational complexity.And,the usage of GPU in matrix calculation further accelerates the raw data generation.The experimental results validate both the efficiency and accuracy of this SAR raw data simulator are high.
What problem does this paper attempt to address?