超宽带混合极化探地雷达系统及目标特性反演成像方法研究

An ultra-wideband (UWB) hybrid polarimetric ground penetrating radar (GPR) system is going to be developed based on a vector network analyzer and a hybrid polarimetric antenna array, which consists of a circularly polarized transmitting antenna and two linearly polarized receiving antennas. The two receiving antennas are orthogonally oriented and can record the radar reflection signal simultaneous. Therefore, the two receiving channels are completely coherent. .Base on the hybrid polarization system, we are going to develop a vector migration algorithm, a polarization decomposition algorithm and a polarimetric attribute analysis method to characterize and extract the geometric features of the subsurface targets, e.g. shape, size, depth and etc. Furthermore, we are going to develop a full waveform inversion algorithm, which make full use of the hybrid polarimetric GPR data, to obtain the physical features of the subsurface targets and its surrounding media, i.e. the dielectric permittivity and electric conductivity. .To eliminate the influence of antenna coupling and radiation pattern on the imaging resolution, we are going to build up a green function table, which is calculated from a complete model consisting of the homogeneous subsurface medium, as well as the hybrid polarimetric antenna array above the ground surface. The green function table would contain the green function in subsurface medium under different operating frequency and when the subsurface medium has different dielectric permittivity and electric conductivity. The full-waveform inversion and imaging algorithm can look for the green function table to eliminate the influence of the antenna coupling and radiation pattern and greatly save the computation cost. The nonlinear inversion and imaging algorithm explores the vector characteristic of the GPR reflection signal and improve the inversion accuracy. .The output of this research project will increase the resolution, accuracy and faith of GPR detection in field and further expand the applications in engineering, environment, geology and so on.

在实验室内基于矢量网络分析仪实现一套采用一个圆极化发射天线和多个线性极化接收天线的超宽带混合极化探地雷达天线阵列系统，各接收通道可同时工作，保证了不同极化雷达反射信号的相干性。基于该系统，采用矢量偏移成像、极化分解和极化属性分析等方法描述和识别地下目标体的形态、尺寸和埋深等几何特性；并采用多极化全波形联合反演方法精确地获取地下目标体的介电常数和电导率等物理特性。为消除天线耦合干扰和辐射方向图的影响，拟建立包含天线和半空间地下介质的仿真模型计算不同工作频率、不同电性参数地下介质中的格林函数列表，为矢量偏移成像和波恩迭代全波形反演算法提供查表服务，实现软件去耦的同时，可大大提高计算效率。基于混合极化的非线性反演成像方法可充分利用地下目标反射信号的矢量特性，提高反演精度。本项目研究成果可提升探地雷达资料的定量解释水平，提高探测精度和成功率，对进一步扩展探地雷达的应用面具有重要的实际意义。

本项目首先搭建了一套由一个螺旋圆极化发射天线和两个线性极化Vivaldi天线所组成的混合紧缩极化天线阵列探地雷达系统，在此基础上提出了一种极化分解方法，可得到全极化雷达数据，并用一个格栅三面角进行了极化校准实验，由此可以比较准确的获取地下目标的极化散射矩阵。其次，研究了探地雷达的逆时偏移成像算法，使用时域有限差分法编制了二维和三维逆时偏移成像程序，并在此基础上，基于分层介质的并矢格林函数提出了一种频域逆时偏移成像算法，可获得与时域逆时偏移成像算法几乎一样的高精度地下目标成像结果，相比时域逆时偏移成像算法，计算效率可提高2-3个数量级。再次，提出了一种基于目标极化散射矩阵和Alford极化旋转的地下线性目标走向估计算法，在沙坑内开展了室内模拟试验，证明了该算法的有效性和高精度。然后，利用极化探地雷达开展了木结构房屋墙体的地震损伤的探测实验，证明极化雷达结合偏移成像可对墙面装饰材料下部木结构微小变形进行精细探测。最后，将逆时偏移成像技术应用到我国嫦娥五号月壤结构雷达探测仪的地面验证试验中，实现了嫦娥五号着陆器复杂电磁环境中空耦雷达数据的高分辨率成像，火山灰坑内试验结果表明，应用逆时偏移成像技术可在三小时内实现对嫦娥五号着陆器钻头下方2m深度内的月壤结构和月岩进行高分辨率成像，分辨率优于3 cm。