Automatic Evaluation of Pavement Thickness in GPR Data with Artificial Neural Networks

Abstract

The ground penetrating radar (GPR) is one of the most frequently recommended non-destructive methods for the pavement thickness measurement. Due to the rapid growth of GPR data in the recent years, the development of automatic data processing techniques is required. In this paper we propose to use one type of artificial neural network, the multilayer perceptron (MLP), for automatic selection of the pavement boundaries. The experimental results indicate that machine learning techniques can be used for robust road structure evaluation.