15-01-2013, 02:04 PM
GROUND PENETRATING IMPULSE RADAR FOR LANDMINE DETECTION
[attachment=46970]
ABSTRACT
The video impulse ground penetrating radar (GPR) system
for detection of small and shallow buried objects has been
developed. The hardware combines commercially available
components with components (e.g. antennas) specially
developed or modified for being used in the system. The
GPR system has been designed to measure accurately
electromagnetic field backscattered from subsurface targets
in order to allow identification of detected targets through
solution of the inverse scattering problem. The GPR has
been tested in different environmental conditions and has
proved its ability to detect small and shallow buried targets.
Key words: ground penetrating radar, video impulse system,
landmine detection, ultra-wideband antenna system.
INTRODUCTION
Recently considerable efforts are put in development of
GPR systems for detection of surface-laid and shallow
buried targets such as antipersonnel landmines. Two crucial
demands to any GPR system for landmine detection are
99.6% probability of detection and low false alarm rate.
While detectability of the system can be improved by means
of improving the resolution and the dynamic range of a
system, decrease of the false alarm rate can be achieved only
via localization, classification and identification of detected
targets. Solution of the latter problem requires accurate
measurements of the electromagnetic field scattered from
the subsurface. This qualitatively new demand makes the
principal difference between usual GPR and GPR for
landmine detection: the first one should just detect the field
scattered from a buried target (i.e. distinguish this field from
all other electromagnetic fields) while the second one should
measure accurately the scattered field (i.e. determine
magnitude of the field as a function of time). From the
measured values of the scattered field different inverse
scattering methods can determine localization, size, shape
and even spatial distribution of dielectric permittivity within
the buried target.
Having in mind this principal difference the International
Research Centre for Telecommunications-transmission and
Radar (IRCTR) in the Delft University of Technology has
developed two GPR systems dedicated to landmine
detection: a video impulse system and a stepped-frequency
continuous wave system. In this paper the main guidelines
of the video impulse system design are presented. The
stepped-frequency system is described elsewhere.
HARDWARE DESCRIPTION
The impulse GPR system developed in IRCTR for landmine
detection comprises a pulse generator, an antenna system, a
signal conditioner and a sampling converter. The pulse
generator delivered by SATIS Co. produces 0.8ns
monocycle. The unique feature of this generator is its small
trailing oscillations, which are below 2.4% of the maximum
amplitude during the first 2ns and below 0.5% afterwards
(Fig. 1). The advantage of a monocycle in comparison with
a monopulse is that the frequency spectrum of the first one
decreases to zero at low frequencies, which cannot be
efficiently transmitted via the antenna system, while the
frequency spectrum of the second one has a global
maximum there. As a result, the magnitude of the field
radiated by an antenna system fed by a monocycle is
considerably larger than the magnitude of the field radiated
by the antenna system fed by a monopulse with the same
magnitude. The generator spectrum covers a wide frequency
band from 500MHz till 2GHz on 3dB level. At frequencies
below 1GHz, attenuation losses in the ground are small
(Daniels, 1996) and considerable penetration depth can be
achieved. However, landmine detection requires down-range
resolution (in the ground) of the order of several
centimeters, which can be achieved using frequencies above
1GHz. It was found experimentally that the 0.8ns monocycle
satisfies penetration and resolution requirements. The
spectrum of this pulse (Fig. 2) has a maximum at
frequencies where the attenuation losses in the ground start
to increase. So the spectral content of the monocycle below
this maximum penetrates deep into the ground and the
spectral content above this maximum provides sufficient
down-range resolution.
ANTENNA SYSTEM
The antenna system is one of the most critical parts of every
GPR, because its performance depends strongly on the
antenna system. The antenna system should satisfy a number
of (sometimes contradictory) demands. The transmit antenna
should:
1. radiate short ultra-wideband (UWB) pulse with small
ringing;
2. radiate electromagnetic energy within a narrow cone in
order to filter out undesirable backscattering from
surrounding objects;
3. produce an optimal footprint on the ground surface and
below it (size of the footprint should be large enough
for SAR processing but at the same time it should be
small enough to reduce surface clutter);
4. the waveform of the radiated field on the surface and in
the ground should be the same;
5. the waveform of the radiated field in the ground should
not depend on type of the ground (i.e. its dielectric
permittivity);
In order to allow successful SAR processing for the given
frequency band the scattered field should be measured with
a cross-range step 3cm maximum. Besides the measurement
plane should be sufficiently elevated above the ground
surface in order to avoid influence of evanescent fields.
Together with operational demands for landmine detection it
means elevation of the receive antenna at least 10cm above
the ground. Thus the receive antenna should:
1. receive the field in a local point (effective aperture
should not be larger than 1cm2);
2. provide sufficient sensitivity in order to receive very
weak fields;
3. be elevated at least 10cm above the ground surface;
4. allow time windowing to isolate the direct air wave
from the ground reflection.
Additionally a possibility to measure simultaneously
backscattered field in two orthogonal polarizations is
desirable.
To satisfy demands NN.3-5 for the transmit antenna it was
decided to implement the far-field approach, meaning that
the transmit antenna is elevated sufficiently high above the
ground. The other demands NN.1-2 can be satisfied by
using a transient antenna with reasonably high directivity.
Such antennas are not commercially available and design of
such an antenna is extremely difficult. In close collaboration
with SATIS Co. a dielectric filled TEM horn (DTEM) has
been designed (Yarovoy, Schukin and Ligthart, 2000),
which is ultrawideband, has linear phase characteristics over
the whole operating frequency band, has constant
polarization and possesses short ringing. The waveform of
the electric field radiated from this antenna fed by the 0.8ns
monocycle generator is presented in Fig. 3 and the antenna
footprint at a distance 54cm from the antenna aperture is
presented in Fig. 4. It can be seen that the footprint at 3dB
level has an elliptic shape with halfaxes 21cm and 27cm.
The waveform of the radiated field remains the same within
the whole footprint (on 3dB level).
CONCLUSION
The video impulse ground penetrating radar system for
detection and identification of small and shallow buried
objects has been developed in IRCTR. First experimental
results show that the system can detect small dielectric and
metal targets at a depth up to 50cm. In the next step of the
program software for image processing, localization and
identification of targets will be developed and implemented
into the system.
[attachment=46970]
ABSTRACT
The video impulse ground penetrating radar (GPR) system
for detection of small and shallow buried objects has been
developed. The hardware combines commercially available
components with components (e.g. antennas) specially
developed or modified for being used in the system. The
GPR system has been designed to measure accurately
electromagnetic field backscattered from subsurface targets
in order to allow identification of detected targets through
solution of the inverse scattering problem. The GPR has
been tested in different environmental conditions and has
proved its ability to detect small and shallow buried targets.
Key words: ground penetrating radar, video impulse system,
landmine detection, ultra-wideband antenna system.
INTRODUCTION
Recently considerable efforts are put in development of
GPR systems for detection of surface-laid and shallow
buried targets such as antipersonnel landmines. Two crucial
demands to any GPR system for landmine detection are
99.6% probability of detection and low false alarm rate.
While detectability of the system can be improved by means
of improving the resolution and the dynamic range of a
system, decrease of the false alarm rate can be achieved only
via localization, classification and identification of detected
targets. Solution of the latter problem requires accurate
measurements of the electromagnetic field scattered from
the subsurface. This qualitatively new demand makes the
principal difference between usual GPR and GPR for
landmine detection: the first one should just detect the field
scattered from a buried target (i.e. distinguish this field from
all other electromagnetic fields) while the second one should
measure accurately the scattered field (i.e. determine
magnitude of the field as a function of time). From the
measured values of the scattered field different inverse
scattering methods can determine localization, size, shape
and even spatial distribution of dielectric permittivity within
the buried target.
Having in mind this principal difference the International
Research Centre for Telecommunications-transmission and
Radar (IRCTR) in the Delft University of Technology has
developed two GPR systems dedicated to landmine
detection: a video impulse system and a stepped-frequency
continuous wave system. In this paper the main guidelines
of the video impulse system design are presented. The
stepped-frequency system is described elsewhere.
HARDWARE DESCRIPTION
The impulse GPR system developed in IRCTR for landmine
detection comprises a pulse generator, an antenna system, a
signal conditioner and a sampling converter. The pulse
generator delivered by SATIS Co. produces 0.8ns
monocycle. The unique feature of this generator is its small
trailing oscillations, which are below 2.4% of the maximum
amplitude during the first 2ns and below 0.5% afterwards
(Fig. 1). The advantage of a monocycle in comparison with
a monopulse is that the frequency spectrum of the first one
decreases to zero at low frequencies, which cannot be
efficiently transmitted via the antenna system, while the
frequency spectrum of the second one has a global
maximum there. As a result, the magnitude of the field
radiated by an antenna system fed by a monocycle is
considerably larger than the magnitude of the field radiated
by the antenna system fed by a monopulse with the same
magnitude. The generator spectrum covers a wide frequency
band from 500MHz till 2GHz on 3dB level. At frequencies
below 1GHz, attenuation losses in the ground are small
(Daniels, 1996) and considerable penetration depth can be
achieved. However, landmine detection requires down-range
resolution (in the ground) of the order of several
centimeters, which can be achieved using frequencies above
1GHz. It was found experimentally that the 0.8ns monocycle
satisfies penetration and resolution requirements. The
spectrum of this pulse (Fig. 2) has a maximum at
frequencies where the attenuation losses in the ground start
to increase. So the spectral content of the monocycle below
this maximum penetrates deep into the ground and the
spectral content above this maximum provides sufficient
down-range resolution.
ANTENNA SYSTEM
The antenna system is one of the most critical parts of every
GPR, because its performance depends strongly on the
antenna system. The antenna system should satisfy a number
of (sometimes contradictory) demands. The transmit antenna
should:
1. radiate short ultra-wideband (UWB) pulse with small
ringing;
2. radiate electromagnetic energy within a narrow cone in
order to filter out undesirable backscattering from
surrounding objects;
3. produce an optimal footprint on the ground surface and
below it (size of the footprint should be large enough
for SAR processing but at the same time it should be
small enough to reduce surface clutter);
4. the waveform of the radiated field on the surface and in
the ground should be the same;
5. the waveform of the radiated field in the ground should
not depend on type of the ground (i.e. its dielectric
permittivity);
In order to allow successful SAR processing for the given
frequency band the scattered field should be measured with
a cross-range step 3cm maximum. Besides the measurement
plane should be sufficiently elevated above the ground
surface in order to avoid influence of evanescent fields.
Together with operational demands for landmine detection it
means elevation of the receive antenna at least 10cm above
the ground. Thus the receive antenna should:
1. receive the field in a local point (effective aperture
should not be larger than 1cm2);
2. provide sufficient sensitivity in order to receive very
weak fields;
3. be elevated at least 10cm above the ground surface;
4. allow time windowing to isolate the direct air wave
from the ground reflection.
Additionally a possibility to measure simultaneously
backscattered field in two orthogonal polarizations is
desirable.
To satisfy demands NN.3-5 for the transmit antenna it was
decided to implement the far-field approach, meaning that
the transmit antenna is elevated sufficiently high above the
ground. The other demands NN.1-2 can be satisfied by
using a transient antenna with reasonably high directivity.
Such antennas are not commercially available and design of
such an antenna is extremely difficult. In close collaboration
with SATIS Co. a dielectric filled TEM horn (DTEM) has
been designed (Yarovoy, Schukin and Ligthart, 2000),
which is ultrawideband, has linear phase characteristics over
the whole operating frequency band, has constant
polarization and possesses short ringing. The waveform of
the electric field radiated from this antenna fed by the 0.8ns
monocycle generator is presented in Fig. 3 and the antenna
footprint at a distance 54cm from the antenna aperture is
presented in Fig. 4. It can be seen that the footprint at 3dB
level has an elliptic shape with halfaxes 21cm and 27cm.
The waveform of the radiated field remains the same within
the whole footprint (on 3dB level).
CONCLUSION
The video impulse ground penetrating radar system for
detection and identification of small and shallow buried
objects has been developed in IRCTR. First experimental
results show that the system can detect small dielectric and
metal targets at a depth up to 50cm. In the next step of the
program software for image processing, localization and
identification of targets will be developed and implemented
into the system.