Wavelet-based method for detection of shallowly buried objects from GPR data

A method for simultaneous clutter suppression and signal detection in ground penetrating radar (GPR) data is presented. It computes a running estimate of the background signal and detects abrupt changes from the estimated signal, assumed to correspond to the returns from a buried object, using a translation invariant (TI) wavelet packet decomposition. The detection statistic is defined as the sum of several peak energy values of the TI wavelet representation and is compared to an adaptively determined threshold. The technique performed satisfactorily when tested on a number of targets buried in several different soil environments.