Toward true amplitude processing of GPR data

For proper imaging and inversion, four parameters need to be known accurately. These are position of the transmitter and receiver antennas, the wave velocity distribution in the medium, the polarization and the amplitude characteristics of the antennas used. The position information problem is solved by introducing an automated acquisition frame with high accuracy in horizontal and vertical positioning of the antennas over a fixed area. For object detection and characterization purposes, we use elevated antennas, which implies that at least a two medium background medium must be used for accurate velocity information. The velocity in air is known, while that of the Earth must be obtained from independent measurements. The polarization and amplitude characteristics can be obtained for elevated antennas by modeling the transmitter and receiver and validate the model with measurements in air. When the antenna model is accurate, it is possible and necessary to take direct coupling with the ground into account for surface laid antennas. In general, this leads to formulations of the direct source and receiver problems, which can be solved locally by finite element techniques or by global techniques such as integral equation formulations. For elevated antennas, or when time windowing is possible for surface laid antennas, a second way of obtaining information on the wave field, both polarization and amplitude, that is emitted into the ground is to record it in a certain plane in the air. In a configuration with constant velocity horizontal layers, only the down going wave field in the upper half space below the antennas, contributes to the down going wave field in the layered Earth. Accurate knowledge of this down-going wave-field allows the wave-field extrapolation into the layered Earth, which technique can be used in multi-component imaging algorithms accounting for polarization and amplitude. Here the effects of knowing the four parameters are shown. It demonstrated t- at the multi-component, or full polarimetric, imaging technique is a tool of full operational value only when all four parameters are accurately known. Only then the required increased acquisition effort pays off.