Multi-Feature Based Multiple Landmine Detection Using Ground Penetration Radar

Abstract

This paper presents a novel method for detection of multiple landmines using a ground penetrating radar (GPR). Conventional algorithms mainly focus on detection of a single landmine, which cannot linearly extend to the multiple landmine case. The proposed algorithm is composed of four steps; estimation of the number of multiple objects buried in the ground, isolation of each object, feature extraction and detection of landmines. The number of objects in the GPR signal is estimated by using the energy projection method. Then signals for the objects are extracted by using the symmetry filtering method. Each signal is then processed for features, which are given as input to the support vector machine (SVM) for landmine detection. Three landmines buried in various ground conditions are considered for the test of the proposed method. They demonstrate that the proposed method can successfully detect multiple landmines.