大场景下地海环境与目标的复合电磁散射建模方法研究

In the field of microwave remote sensing, the extraction of target signal becomes more difficult due to the effect of electromagnetic (EM) scattering wave from the environment. Thus, the study on the composite scattering from targets in complex environment will be helpful to improve the ability of target recognition. On the other hand, a wideband radar signal is needed to acquire a well resolution capability for more detailed information of the interested targets. However, the wideband signal contains more frequency components, and it leads to a heavier computing burden in the simulation of radar echo. To meet the requirement of engineering application, the combination of the hybrid method of geometry optics and physical optics method (GO/PO) and the small slope approximation (SSA) is used to calculate the composite scattering properties of targets in complex environments. To analyze the composite EM scattering of complex targets over large rough sea surface on personal computer with high efficiency, the spectral decomposition modeling method is proposed to solve the problem of size restriction of rough surface caused by the large memory costs when using the SSA method. To improve the efficiency, the rectangular wave beam based GO/PO method is used to accelerate the process of ray tracing. In the final, to meet the requirement of high efficiency in the simulation of composite EM scattering properties of large scene under a frequency-sweep mode, the compute unified device architecture (CUDA) is employed to further speed up the EM scattering calculation on display card.

在微波遥感领域中，由于受环境电磁散射杂波的影响，从电磁散射回波中提取目标信号难度将变大，因此研究环境与目标的复合电磁散射机制将有助于提高目标识别能力。此外，为提高雷达的分辨率探测物体更清楚的细节信息，需要增大信号的带宽。然而更宽的信号带宽则意味着更多的频率成份，使得模拟其回波的计算量更大。为满足工程应用要求，本项目采用几何光学与物理光学(GO/PO)和小斜率近似(SSA)混合算法研究粗糙面与目标的复合电磁散射特性。为了在个人计算机上分析符合实际情况的大场景目标与粗糙面复合电磁散射特性，采用了谱分解建模方法处理粗糙面的散射问题，解决了小斜率近似占用大量内存造成计算场景尺寸受限问题。为提高计算效率，采用了基于矩形波束的GO/PO算法，可以加速射线追踪过程。 最后利用通用并行计算架构 (CUDA) 技术，在显卡中加速电磁散射计算，可以进一步提高计算效率，满足在大场景电磁仿真中快速计算扫频电磁散射特性的要求。

为了分析大场景目标与环境复合场景电磁散射特性并满足实际工程应用的需求，本项目从精度和效率两方面开展复合场景的散射模型研究。首先，采用几何光学与物理光学(GO/PO)和小斜率近似(SSA)混合算法，考虑多次散射作用的效果，可以较准确的分析复合场景的电磁散射特性。其次，采用了基于矩形波束的GO/PO算法，可以加速射线追踪过程。最后，为了进一步提高计算效率，利用NVIDIA推出的通用并行计算架构(CUDA)，在显卡中实现算法的并行加速，满足工作应用对计算效率的要求。有了高效的复合场景电磁散射模型，就可以快速分析其扫频特性，结合宽带信号与合成孔径雷达(SAR)的照射过程，就可以得到符合实际情形的SAR图像，为目标识别或是隐身研究提供理论依据和仿真数据。