D.E. Edemskiy, F.D. Edemskiy, P.A. Morozov

In this article we consider a problem of subsurface medium properties estimation (permittivity and layer thickness) on GPR data acquired along the trace during cross-section survey. Carrying out a cross-section of the survey area is the most widespread method of GPR data acquisition, as it does not take much time and it is very informative. Signal waveforms acquired by this method form a time-section of the survey trace. However, usually there is no way for cross-section to provide an operator with any justified information on values of subsurface layers properties without any a priori information, due to an ambiguous interpretation of a signal delay dependence on seam depths and speed of wave propagation. So called common mid point (CMP) method partly solves this problem. Using the CMP method in the surveys allows one to localize horizontal bedding interfaces and identify them. Usually CMP data is acquired in a single characteristic point of a survey trace, where there is a clear distinction between the layers, and the results of the CMP data analysis are extrapolated to the whole survey area in case of flat-layered medium. The use of the CMP method in all points of a trace piles up working time too much. The main point of the proposed method is carrying out two cross-sections along the same survey trace with two fixed and different antenna separations. It increases working time twice as much, but it makes possible to estimate some properties of subsurface geological horizons along the whole survey trace, because only two points of time-path curve are needed for relative permittivity and layer thickness estimation. Thus operator can observe changes of layers - moisture and thickness. To carry out an experiment two areas were chosen. They had different properties of underlying surface and moisture of subsurface layers. A survey using the CMP method was conducted to identify bedding interfaces and to compare with the processing results on the data, acquired using the proposed approach. Analysis of the results showed that not any couple of points of the experimental time-path curve is acceptable for soil properties estimation. Too large antennas separation leads to difficulties for reflected signal identification, because it becomes weak and subjected to strong distortion by clutter and local objects reflections. It is quite helpful to use the CMP method for signal to noise ratio estimation and optimal antenna separation determination to minimize clutter impact on the results. As for minimal of technically possible antenna separations, practice showed that it leads to excessive local objects sensitivity, like, for instance, stones and various rubbish. In addition, front ends of the device and hence results are subjected to strong antennas mutual influence that happens in case they are placed too close to each other. The results, which were obtained by this method, are subjected to different distortions and quite careful acquisition approach is required. In spite of the CMP method being formally unfounded in case of sloping bedding interfaces, errors caused by this effect usually are not strong what makes the proposed method applicable in most situations. Application of this method will simplify the process of geological profile estimation what makes GPR surveys more useful in engi-neering surveys. This approach is available for all types of GPRs with separable transmitter and receiver antennas.