Automatic detection of asphalt layer thickness based on Ground Penetrating Radar

Abstract

In road evaluation, Ground Penetrating Radar (GPR), which is a Non-destructive Testing (NDT) technique, has been commonly adopted to estimate asphalt layer thickness of the pavement. This study focuses on implementing a pixel-wise detection approach to automatically detect asphalt layer of the pavement from the acquired GPR images. The dielectric properties of the pavement compositions are linked to intensity levels in GPR images. The interface between asphalt layer and base layer below it can be tracked and located continuously. Thus, asphalt layer thickness for each location can be estimated based on the calibrated spatial resolution of the GPR images. Ten GPR images taken from ten different roads are used for experiments. The road distance recorded in these GPR images ranges from 245m to 2842m. The validity of the automatic detection approach is evaluated by comparing the results with the manual detection results of an expert. The accuracy is quantified by computing the absolute difference between the automatic thickness estimation and the manual thickness measurement. The results indicate that the proposed approach is able to locate asphalt layer and measure its thickness accurately. The accuracy is up to 3.14mm when compared with the manual detection.