Numerical Parametric Study: Effect of GPR Source on MPM-Inversion Performances

This paper presents a parametric study based on ground penetrating radar signals inversion, using matrix pencil method for improving detection of buried utility networks in urban context. This study shows negligible impact and highlighting that using calibrated signals (object signature only) instead of raw ones (source + object signature) can simplify database design (size and time calculation) for machine learning-based hybridization.