Detection of debondings with ground penetrating radar using a machine learning method
TODKAR ; LE BASTARD ; IHAMOUTEN ; BALTAZART ; DEROBERT ; FAUCHARD ; GUILBERT ; BOSC
Type de document
COMMUNICATION AVEC ACTES INTERNATIONAL (ACTI)
Langue
anglais
Auteur
TODKAR ; LE BASTARD ; IHAMOUTEN ; BALTAZART ; DEROBERT ; FAUCHARD ; GUILBERT ; BOSC
Résumé / Abstract
In the field of civil engineering, Ground Penetrating Radar (GPR) is the most widely used method of Non-Destructive Testing (NDT). Using supervised learning methods or signal processing methods, it is possible to analyze the sub-surface defects in pavement. In this paper, we propose to use a machine learning method called Support Vector Machines (SVM) to detect the presence of debondings within the pavement. Here, the ground-coupled GPR in quasi mono-static configuration along with SVM is used to detect debondings. The experiments are done on bituminous concrete pavements with various material characteristics. The classification results are good and show the efficiency of the detection process.
