Adaptive linear prediction based buried object detection with varying detector height

In the GPR underground inspection problem, received GPR signal is notably dependent on the height of the detector head. A great deal of the former research in this area assumes constant detector height, as absence of this constancy will result in undesired change in the returning signal, which will reduce the success of the detection and identification stages. In this study, we propose a buried object detection method which removes the deforming effect of the variable sensor height. In our approach, adaptive one-sided linear prediction is utilized in a specific manner. The proposed algorithm is causal, which makes it convenient for real time buried object detection applications.