The development of underground structure estimation system using the electromagnetic wave

利用电磁波进行地下勘探的技术大致分为两类。一是探地雷达（GPR）。在要求无损勘探和评估的情况下，探地雷达已经吸引了人们对探测管道（通信电缆、煤气管道、电缆）、污水管道、隧道和其他地下空间的兴趣。另一种是钻孔雷达。钻孔雷达用于探测岩石裂隙、空洞、地层评价等方面的研究已引起人们的极大兴趣。近年来，地下目标的位置和大小以及其电性（介电常数和电导率）分布都引起了人们的兴趣。提出了一种利用实测时域场数据反演地下圆柱体电性分布的新算法。完成本课题的主要成果如下：（1）改进了自由空间介质目标重建的FBTS方法，使之适用于地下勘探。(2)数值模拟表明，改进的FBTS方法可以估计地下结构的电特性。(3)由于FBTS方法的计算时间大大延长，以扩大搜索区域，通过将合成孔径处理与FBTS方法相结合，实现了计算速度的加速。(4)为了能够从具有大噪声的数据中进行重建，提出了滤波FBTS方法。(5)但是，重建区域的扩大和大噪声的存在降低了重建图像的质量，并且需要很长的CPU时间。为了改进这些算法，我们提出了一种新的重建算法，将滤波FBTS算法与多重网格相结合。

Underground exploration technology using the electromagnetic wave is divided roughly into 2 types. One is the ground penetrating radar (GPR). GPR has been attracting interest for purpose of detecting pipes (communication cables, gas pipes, electricity cables), sewage lines, tunnels and other subsurface spaces in situations calling for nondestructive exploration and evaluation. The other is the borehole radar. The borehole radar has been attracting interest for purpose of detecting the crack of the rock and the cavity, evaluating the stratum. In recent years, not only the locations and sizes but also the electrical property (permittivity and conductivity) distributions of underground objects are of interest. A new inversion algorithm for reconstructing electrical property distributions of underground cylindrical objects by means of the measured time-domain field data is presented. Main results of achieving the term of this project are as follows :(1) FBTS method proposed for reconstructions of dielectric objects in free space is improved in order to apply to the underground exploration.(2) Numerical simulations show that improved FBTS method can estimate the electric properties of underground structure.(3) As the calculation time of FBTS method is very much lengthened for widening a search region, the speedup of the computation speed was carried out by combining the synthetic aperture processing with the FBTS method.(4) The Filtered FBTS method was proposed in order to enable the reconstruction from the data with large noise.(5) However, the expansion of reconstruction area and existence of large noise degrade the quality of reconstructed image and require long CPU time. In order to improve them, we propose a new reconstruction algorithm that combined Filtered FBTS algorithm with multigrid.