Buried threat detection using a handheld ground penetrating radar system

Abstract

In this work, we explore the efficacy of two buried threat detectors on handheld data. The first algorithm is an energy-based algorithm, which computes how anomalous a given A-scan measurement after it is normalized according to its local statistics. It is based on a commonly used prescreener for the Husky Mounted Detection System (HMDS). In the HMDS setting measurements are sampled on a crosstrack-downtrack grid, and sequential measurements are at neighboring downtrack locations. In contrast, in the handheld setting sequential scans are often taken at neighboring crosstrack locations, and neighboring downtrack locations can be hundreds of scans away. In order to include both downtrack and crosstrack information, we compute local statistics over a much larger area than in the HMDS setting. The second algorithm is a shape-based algorithm. Shape Invariant Feature Transform (SIFT) features, which capture the gradient distributions of local patches, are extracted and used to train a non-linear Support Vector Machine (SVM). We found that in terms of AUC, the SIFT-SVM algorithm results in a 2.2% absolute improvement over the energy-based algorithm, with the greatest gains seen at lower false alarm rates.

Estimation of Handheld Ground-Penetrating Radar Antenna Position with Pendulum-Model-Based Extended Kalman Filter

“…Due to a relatively large GPR antenna beamwidth, the same buried objects are illuminated many times from different antenna locations and consequently from different distances. Therefore, the echo signals form hyperbolic structures visible in the B-scans [14,23,39,46,47]. An example of such a structure for a single-point object is shown as a red hyperbole in Figure 2b.…”

“…The GPRs on vehicle platforms, however, have limited utility in difficult terrain: mountainous areas, forests, dumps, urban surfaces covered with debris, or interiors of buildings, where landmines and other explosives can be typically found. A good solution applicable in such areas is a handheld version of the ground penetrating radar (HH-GPR) [39][40][41][42]. The problem of estimating the antenna position of such type of radar is addressed in this paper.…”

“…A single echogram was obtained for only one GPR antenna position with coordinates (i, j) is a one-dimensional signal representation, called an A-scan (Figure 2a). Time delays of the signal peaks in the A-scan are usually converted into respective depths and the Z-axis is scaled in the distance units [21,23,39,46,47]. An analysis of GPR data is typically based on a two-dimensional signal representation, called a B-scan, which is a dataset created from many A-scans acquired for various antenna locations along a usually linear scanning path, as shown in Figure 2b.…”