Ground-penetrating radar (GPR) is a nondestructive investigation tool that is usually used in flexible pavement evaluation to estimate the thicknesses of the various layers composing the pavement. GPR is also used in flexible pavements to detect subsurface distresses, such as moisture accumulation and air voids. For rigid pavements and bridge decks, GPR is used to measure the thickness of the concrete slab and detect the location of reinforcing bars (rebar). Rebar detection is typically achieved, in this case, when an experienced operator finds the rebar's classic parabolic signature in the GPR data. This paper presents image-processing techniques that can be used to detect the rebar parabolic signature automatically in GPR data collected from rigid pavements with a high-frequency ground-coupled antenna. After detection of the rebar, the reflected parabolic shape is fit to a theoretical reflection model to estimate the pavement's dielectric constant and the rebar depth. The algorithms were validated on GPR data collected from a known continuously reinforced concrete pavement section. The technique showed an average error of 2.6% on the estimated rebar cover depth.
A Machine Learning Scheme for Estimating the Diameter of Reinforcing Bars Using Ground Penetrating Radar
