Skip to Main content Skip to Navigation

Imagerie georadar et modelisation des diffractions multiples

Abstract : Ground Penetrating Radar is widely used to prospect the near surface with high frequency electromagnetic
waves (F> 10 MHz). Imaging the dielectric contrasts in the first meters of the earth shows the changes in lithology, fracturation, or water content... In these dielectric media the electric field induces a polarisation, which results in the distortion of the signal. Following precedent studies, we use a frequency power law for the electric susceptibility to model this dispersion effect.
With such a sensibility to the different media and a resolution ranging from 5 centimeters to several
decimeters, the strong heterogeneity of the near surface needs to be considered for an efficient data processing. For this reason, we developed a new modeling algorithm using the finite difference time domain scheme and proposed a modified recursive convolution scheme to include all type of dispersive properties.
As boundary conditions, we implemented absorbing perfect match layers in an original way.
We also studied the reflexion coefficients inside aeolian sand dunes and concluded that they can be
explained by small variations in the water content. The analysis of the spectral ratio between signals at
different offsets provides us with a new tool to estimate the dielectric permittivity. These data allow us to model more realistic synthetic cases. At last, we successfully used the same modeling program, but
with time reversal, to propagate the recorded fields back to the source.
Complete list of metadata

Cited literature [4 references]  Display  Hide  Download
Contributor : Jean-Francois Girard <>
Submitted on : Friday, March 9, 2007 - 5:39:28 PM
Last modification on : Thursday, April 23, 2020 - 2:26:29 PM
Long-term archiving on: : Friday, September 21, 2012 - 12:45:17 PM


  • HAL Id : tel-00135945, version 1



Jean-Francois Girard. Imagerie georadar et modelisation des diffractions multiples. Géophysique [physics.geo-ph]. Université Louis Pasteur - Strasbourg I, 2002. Français. ⟨tel-00135945⟩



Record views


Files downloads