基于电磁仿真的粗糙面上二维目标参数反演

针对海洋(或陆地)环境中目标电磁逆散射问题开展研究,建立了粗糙面上简单导体目标的电磁散射模型,通过精确的电磁散射建模得到散射回波数据,包括空间分布采样和频率采样,再分别应用BP神经网络(BPNN)和支持向量机(SVM)两种方法反演粗糙面上二维导体目标的特征几何尺寸.在粗糙面上圆柱几何参数反演以及海上舰船尺寸反演过程中需要考虑粗糙面的随机性,因此对散射数据进行了多次统计,实现了随机粗糙面上方目标的特征参数重构.

Parameters Extraction of the Two-Dimensional Object above or on a Rough Surface Based on the Electromagnetic Simulation

In this paper, the geometric characterization of an object above or on a rough surface is successfully estimated from the simulated back scattered data by the electromagnetic inversion scheme. The inverse model is built based on the EM simulation data computed with a fast forward algorithm considering the multi realization of the random rough surface. Some are desired parameters, the location and the size,of a two-dimensional (2-D) target, like a cylinder above a random rough or a ship on the sea surface is reconstructed by the back propagation neural network (BPNN) and the support vector machine (SVM) respectively.