Sand moisture assessment using instantaneous phase information in ground penetrating radar data Yu Zhang, Dylan Burns, Dryver Huston, Tian Xia School of Engineering, University of Vermont, 33 Colchester Ave, Burlington, VT 05405, USA ABSTRACT In this paper, a method using the instantaneous phase information of the reflection ground penetrating radar (GPR) signal to detect the variation of sand moisture is developed. The moisture changes the permittivity of the medium, which results in different speed when the GPR electromagnetic (EM) wave propagates in the medium. In accordance to this principle, we develop an analytical method to extract GPR reflection signal’s instantaneous phase parameters utilizing Hilbert Transform for sand moisture characterization. For test evaluation, Finite Difference Time Domain (FDTD) numerical simulations using a 3 rd party open source program GprMax V2.0, and laboratory experiments on sand samples are conducted using a commercial GPR (2.3 GHz Mala CX) as the data acquisition system. Keywords: Ground penetrating radar, instantaneous phase, Hilbert transform, sand moisture content, dielectric constant, FDTD simulation 1. INTRODUCTION Saturated sand region in road subsurface layer is a sign of the early stage of roadway sinkhole formation, railroad ballast contamination and sub-base degradation. Non-destructive test of the subsurface saturated region can prevent safety hazards and costly emergency repair on roadway and railway. GPR is a major subsurface imaging tool for non-destructive infrastructure inspection [1]. To characterize the subsurface moisture content, the approach of GPR backscattering signal amplitude measurement has been extensively utilized [2-5]. For EM wave transmitting through the subsurface medium of different moisture levels, the signal attenuation factor is different. Hence, the backscattering amplitude implies different levels of moisture. However, to derive accurate quantitative relationship between the backscattering signal amplitude and the medium moisture (through dielectric property characterization) is very complicated. As shown in Figure 1, the GPR signal is emitted from the transmitter antenna into the subsurface medium and the receive antenna collects the backscattering signals. For the study in this paper, the top layer is the sand layer with various moisture contents and the base layer is the dry concrete.  1 is the echo from the air-sand interface,  2 is the echo from the interface between the sand layer and the concrete layer, and  3 is the echo from the bottom of the concrete layer. The echo signal  2 is measured to monitor the moisture variations in the sand layer. To accurately measure  2 amplitude is not trivial, as it can be easily contaminated by noise, interference and other surface layer reflection signals. Figure 1 GPR detection scheme