Discrimination of Near-Surface Targets in Step-Frequency Radar

An improved frequency domain method for detecting and discriminating shallow subsurface targets is presented. Conventional signal processing technique in step-frequency ground penetrating radar (GPR) is based on inverse fast Fourier transform (IFFT) of reflection coefficient. However, the synthetic range profile generated by IFFT often cannot be used to detect shallowly buried objects due to lack of resolution. Simple modification of the expression for the synthetic range profile that takes into account both amplitude and phase of IFFT results in enhanced resolution and allows discriminating targets according to their reflectivity phases. Further improvement of the performance of the method has been achieved by extracting the reflection from the ground surface. To this end, radar system has been calibrated using optimization procedure, where input data are reference signals measured with the antenna positioned at different distances from metal plate. Free space has been used as another calibration standard. After calibration, the ground surface response can be easily subtracted revealing shallow objects. Experimental examples are given to illustrate the method performance.