Time-frequency domain signature analysis of GPR data for landmine identification

In this paper, the problem of detecting buried antipersonnel (AP) landmines is tackled in the broader context of target identification: determining relevant features, extracted from impulse ground-penetrating radar (GPR) signals, which can be used to classify landmines. These features are extracted in the time-frequency domain using the Wigner-Ville distribution (WVD) and the wavelet transform (WT). Radar data are collected using the MINEHOUNDTM hand-held dual-sensor system over two types of soil and for different landmines and objects. The Wilk's lambda value is used as a criterion for optimal discrimination. Results show that time-frequency signatures from WVD contain more valuable information than the features extracted using WT. Therefore, they could improve landmine and false alarm classification and help to differentiate between two different landmines.