Imaging and inversion of buried objects using GPR

Abstract

The problem of detection of a buried object is formulated as an electromagnetic inverse scattering problem, using algorithms of increasing complexity. Our first approach consists of a back-projection of the measured data and this imaging algorithm enables the location of the suspicious areas, where further investigation is needed. The second approach is effective inversion, where a domain of interest is selected and an effective permittivity and conductivity inside the domain is computed using either a linear or a nonlinear optimization scheme. In our third approach we aim to invert the spatial distribution of the permittivity and conductivity inside the domain of interest. We employ either a linearized inversion scheme based on the Born approximation or a full non-linearized inverse method based on the contrast source inversion (CSI) method.