UXO Detection at Jefferson Proving Ground
Using Ground-Penetrating Radar
STEVEN A. ARCONE, KEVIN O'NEILL, ALLAN J. DELANEY,
AND PAUL V. SELLMANN
INTRODUCTION
trical properties upon the GPR survey parameters of
frequency, towing speed, and data acquisition rate. We
The remote detection of unexploded ordnance
surveyed a site where the types, locations, depths, and
(UXO) has generally relied on the electromagnetic in-
approximate orientations of both ordnance (metallic)
duction (EMI) and ground-penetrating radar (GPR)
and nonordnance (metallic and nonmetallic) are docu-
methods. EMI, the most common method, finds targets
mented. We used supplied location maps to identify the
that contrast strongly in electrical conductivity with their
probable nature of targets detected. The specified depths
host soil by detecting 10- to 40-kHz magnetic fields
are rarely stated to exceed one meter. We used a com-
induced from the target. EMI systems are mobile and
mercially available GPR system and towed 100-, 300-
do not require ground contact, and their performance is
not degraded by minor variations in topography. How-
along established lines. The higher two frequencies pro-
ever, they have little ability to detect nonmetallic tar-
vide high resolution of subsurface objects and are within
gets. GPR, which generally operates between about 50
military specification for ordnance detection, and the
and 1000 MHz, is sensitive to contrasts in both con-
antennas are small and mobile. We used the signal char-
ductivity and dielectric permittivity. Thus GPR can
acteristics obtained from ordnance, nonordnance, and
detect plastic, glass, concrete, or wood when their per-
objects we emplaced to determine soil permittivity and
mittivity contrasts with that of a moist soil. In addition,
diffraction waveforms.
GPR shows the continuity of soil horizons and thus can
Three-dimensional numerical modeling employed an
detect burials of foreign targets by the disturbance of
integral method that allowed incorporation of a reason-
those horizons. GPR performance requires good an-
tennaground contact and a complex soil permittivity
well as arbitrary target shape and orientation. Key ide-
and conductivity that does not rapidly attenuate
alizations were the assumption of homogeneous ground
radiowave energy. The 50- to 300-MHz range usually
and exclusion of ground surface effects. The model was
provides optimum penetration for a given soil conduc-
used to compute fully polarimetric mono- and bistatic
tivity and permittivity. In this report we discuss GPR
scattering over a broad band, for positions directly over
results from a site within the Jefferson Proving Ground
a target as well as along a continuous transect passing
(JPG), where ordnance is buried within a highly attenu-
by the target.
ating soil and for which reports of GPR performance
(NAVEOD 1997) are negative.
SITE DESCRIPTION
Our objectives were 1) to assess the ability of GPR
to detect UXO-type targets at JPG, and 2) to under-
stand the nature and causes of target responses to fos-
JPG (Fig. 1) is located in southern Indiana. Our in-
vestigations were performed within lot 54, known as
ter better target discrimination or explain difficulties
the 40-acre site (Fig. 1). The vegetation at this site is
therein. We were interested in the effect of the soil elec-
Previous Page
