Development of an Efficient and Accurate Forward Model for GPR Imaging of Buried Objects

Imaging buried objects using ground penetrating radar (GPR) has been a key research focus over the years. In GPR imaging, popular migration methods like Kirchhoff's migration (KM) and Frequency-Wavenumber (F-K) migration are used for homogeneous media. They cannot adapt to scenarios like layered media, where electromagnetic (EM) wave velocity changes across layers. Model-based imaging methods are better suited for layered media. Numerical models like finite difference time domain (FDTD) method and analytical models based on dyadic Green's function have been used for subsurface object imaging. However, they are computationally expensive. In this work, a new model based on the full wave model (FWM) of transmitted EM wave component has been developed to represent layered media. The proposed model has been validated with well known FDTD based FWM available in literature. Then time reversal imaging (TRI) method has been implemented on experimental data to perform imaging of buried objects. Peak signal-to-ratio (PSNR) values are evaluated for a qualitative study. The results demonstrate that the proposed model provides the same quality in imaging with enormous computational efficiency compared to the FWM.