Automation of contrast media boundary detection when processing GPR data

Abstract The paper presents the field results of the inspection of transport facilities, in which the contrasting media in terms of radio wave propagation velocities are found. The possibility of high-precision determination of interface boundaries for such media at shallow depths (for example, the contact of asphalt concrete pavement and crushed stone and sand mixture), as well as at high depths (for example, the determination of the bottom of water bodies) is discussed. In addition, the possibility of using artificial neural networks to identify contrasting boundaries on the GPR data and the transition to less contrasting boundaries is justified.