Two-dimensional tomography of crosshole radar data

Ground penetrating radar (GPR) is an electromagnetic method that offers high resolution for near-surface image without damage. Although it is important that the method can identify discontinuities in the medium with high sensitivity, the weakness of the method is that the electromagnetic (EM) wave velocity of the medium is assumed as a constant. The main objective of this thesis is to map the subsurface radar velocity or dielectric coefficient variations in more detail using the crosshole radar method. For this purpose, the inversion process was performed using the first arrival times of the radar wave field from the source point to the receiver point. Within the scope of the thesis, MATLAB and FORTRAN programming languages were used to develope codes for the process steps. In this thesis, firstly, modelling and travel time calculation were performed with the data collection system by creating synthetic datas containing discontinuities of different sizes, velocities and properties that define the environment between the two wells. For the travel time calculation, the two-dimensional (2D) Eikonal equation was solved by second order finite difference approach. In the inversion step, damped least squares and OCCAM inversion techniques were used. In both inverse solutions, the equation was solved by the conjugate gradient technique. It has been shown with synthetic data that the OCCAM inversion method gives better results than the damped least squares solution.