Classification of Buried Land Mines using 3D Templates on GPR Data Sequences

摘要

A series of experiments have been carried out using a ground penetrating radar(GPR) in order to detect and classify anti-tank (AT) and anti-personnel (AP) mine like objects buried in sand. Two different antennas were operated with pulsed or stepped frequency radar emitter. The classification methods is based on 3D templates extracted from data regions known to contain prototype mines. Various modifications of these templates were matched to full data volumes using normalized correlation. Depending on sand moisture, successful recognition was achieved for both AT mines (dry sand sufficient) and AP mines (wet sand).