Two-dimensional imaging and shape recognition of land buried objects through polarimetric ground penetrating radar

The present paper proposes a new ground penetrating radar (GPR) system for obtaining high resolution images for objects buried under the surface of a ground soil. Four sets of clustered two-dimensional data representing the complex scattering parameters (coupling coefficients) between transmitting and receiving antennas are obtained. The paper introduces a new algorithm using a sliding window with a thresholding mechanism to extract a two-dimensional drawing to identify the shape of the buried object from the polarimetric GPR data. The obtained drawing for the buried object shape is compared to a reference target shape and a least-squared-error method is applied to take a match or mismatch decision. The proposed GPR antenna system and the suggested algorithms for shape extraction and target recognition are examined by their application to the scattering parameters data of the four polarization channels obtained through electromagnetic computer simulation using the Finite-Difference-Time-Domain (FDTD) method. A variety of targets with different shapes are modeled and immersed into a ground soil. The numerically obtained results show that proposed GPR system with the suggested algorithms for shape extraction and target shape recognition are successful in all the examined cases.