Effective polarization filtering techniques for ground penetrating radar applications

The effect of different decomposition techniques on the imaging and detection accuracy for polarimetric surface penetrating radar data is studied. We derive the general expressions for coherent polarimetric decomposition using the Stokes parameters and model based polarimetric decomposition using the Yamaguchi technique. These techniques are applied to multi-frequency (0.4-4.8GHz) full polarimetric near-field radar measurements of scattering from surface laid calibration objects and shallow buried landmine types and show in detail how the decomposition results provide effective surface and sub-surface clutter reduction and guide the interpretation of scattering from subsurface objects. Data processing methods assume cross-polar symmetry and a novel bistatic calibration procedure was developed to enforce this condition. The Yamaguchi polarimetric decomposition provides significant clutter reduction and image contrast with some loss in signal power; while Stokes parameters also provide imagery localising targets, complementary information on the scattering mechanism is also obtained. Finally a third novel polarimetric filter was formulated based on differential interferometric polarimetric decomposition. The three combined techniques contribute to a significant improvement of subsurface radar performance and detection image contrast.