ABSTRACT

Landmine detection research has been carried out at the University of Auckland since 1995. Detection technologies currently under investigation include thermal imaging, ground penetrating radar, and explosive sniffing. The thermal imaging work includes investigation of an active system in which the ground is stimulated by the application of microwave energy. The GPR work has concentrated mainly on a swept-frequency radar. The explosive sniffer includes some initial investigation of the quartz microbalance technique. Each area of work will be described and some recent results presented.
 
 

HISTORY

In a presentation at the University of Auckland School of Engineering in 1994, Major John Armstrong of the New Zealand Army Engineers spoke about the work done by the NZ Army in providing mine clearance training in Cambodia and other mine-affected countries. [1] He drew attention to the problem of finding low-metal mines such as the Chinese Type 72, and presented this as a technical challenge to his audience of engineers. [2] In the following year the author responded by assigning a project on landmine detection to two final-year Electrical Engineering students, Mark Le Fevre and Wallace Wong. They looked for ways to distinguish between the mine, and the ground around it. One idea emerged, and took hold: the dry mine was likely to respond differently to microwave radiation, compared to the moist soil. This might produce a temperature effect at the surface, allowing the presence of the mine to be established. [3]

STIMULATED THERMAL IMAGING

In an early test of this idea a thick plastic lid, simulating a mine, was buried in sand in a round open metal container. A microwave oven was used to heat the sand briefly, and then a handheld temperature measuring 'gun' was used to check spot temperatures on the surface. As predicted, a cold spot formed just above where the 'mine' had been buried, revealing its location. Further tests have since been carried out using a thermal imaging camera for a range of conditions and targets. [4] Figure 1 shows the image obtained for a Type 72 apm, 15 mm below the surface of sand, after irradiation for 90 seconds by approximately 700 W of microwave energy at a frequency of 2450 MHz. The resulting image indicates not only the location of the target, but also its size and shape.
 
 

The effectiveness of this method in showing the location of the target is demonstrated when the mine is excavated and a second thermal image taken of the scene (Figure 2). The round hole shows exactly where the mine was, and compares well with the original image on the surface.

In subsequent tests, a microwave source has been mounted on a trolley to allow energy to be distributed fairly evenly over the surface. Tests have been made to determine the locations of buried and completely non-metallic targets (Figure 3a). Figure 3b shows the surface image for the cross-shaped target, and Figure 4 shows the corresponding image after the object has been excavated.
 
 

This technique can certainly be used to determine the position of buried non-metallic objects, but further work is needed to establish its usefulness in a minefield. Issues include power consumption, safety, cost, and effectiveness for deeply-buried objects.

GROUND-PENETRATING RADAR

Many investigators continue to see ground-penetrating radar (GPR) as a possible technology for landmine detection. A difficult problem for all GPR systems is in successfully distinguishing between reflections from the target and the ground, given that anti-personnel mines are often buried shallowly. Target resolution may be improved by operating at higher frequencies, but such frequencies are subject to greater levels of attenuation in the ground.

At the University of Auckland we have developed a computer model which allows an assessment to be made of the likely performance of GPR systems. Variables which can be set include system bandwidth, height of the antenna above ground, depth of mine, thickness of mine, permittivity and conductivity of the mine materials, and type of soil. The model takes account of soil attenuation, and reflections at the various boundaries including the lower surface of the mine. The time-domain output of the model shows the expected received signal, including reflections from the ground surface and the mine.

Some typical results are shown for an X-band (8-12 GHz) radar, with antenna standoff of 25 cm. In order to show best and worst cases, mines are modelled here as either completely metallic, or completely non-metallic. Figure 5 shows the time-domain output for a metallic mine-like target buried at a depth of 5 cm in sand. The reflection from the ground surface is clearly seen, as is the very strong reflection from the metal surface.

Figure 6 shows the expected output for a target made of PVC, other parameters remaining unchanged. The reflection from the target is about 38 dB weaker than before. A reflection from the back of the target may also be seen.

Figure 7 shows the output for the same PVC target in a loamy soil. In this case the target echo is almost obscured by the ground reflection.

In order to test some of our ideas about GPR a swept-frequency (FMCW) system has been developed. A sawtooth frequency sweep is transmitted, and the returning echo is mixed with a fraction of the outgoing signal. The resulting difference frequency, typically at about audio frequency, is a measure of the target range. We have built a prototype radar and used it to get some measurements on buried targets. Figure 8 shows a typical output in the form of a frequency spectrum, with peaks corresponding to reflections from the target, and the ground surface.

ARTIFICIAL OLFACTION

Dogs have been used successfully for some time in landmine detection work. However like humans they get bored, tired, or otherwise disinclined to work. Much work has been done, and is continuing, on producing an 'artificial dog's nose' which could detect the presence of landmines by responding to traces of explosive vapour in the vicinity. Sensitivity needs to be extremely high, of the order of parts per billion. At Auckland we have done some exploratory work on the development of an artificial olfaction system using the quartz crystal microbalance technique, in which a quartz crystal is coated with a substance sensitive to the explosive material. As air is drawn past the crystal, vapour particles interact with the crystal and change its mass slightly. This causes a small but detectable change in the oscillation frequency of the circuit for which the crystal is the frequency-determining element. This work is in its early stages and at the time of writing there are no significant results to report.
 
 

CONCLUSIONS

Three potential technologies for landmine detection are under investigation at the University of Auckland. An artificial olfaction system is in the early stages of development. Some work has been done on ground-penetrating radar, including the development of a computer model and the construction of an FMCW radar. A technique using microwave stimulation and thermal imaging has given encouraging results and is being investigated further.

REFERENCES

1. Engineers for Social Responsibility Newsletter (ISSN 0112-4269), Vol 10, No. 5,September 1994

2. Hidden Killers 1998 - The Global Landmine Crisis, United States Department of State, Office of Humanitarian Demining Programs, Washington DC, September 1998

3. Carter, L.J., Bryant, G.H.B., Le Fevre, M., and Wong, W.C., 'Moisture and landmine detection', Proceedings of EUREL/IEE International Conference on the Detection of Abandoned Landmines (MD96), Edinburgh, IEE Conference Publication Number 431, 83-87, 1996

4. Carter, L.J., O'Sullivan, M.J., Hung, Y.J., and Teng, J.C-C., 'Thermal imaging for landmine detection', Proceedings of IEE International Conference on the Detection of Abandoned Landmines (MD98), Edinburgh, IEE Conference Publication Number 458, 1998
 
 

ACKNOWLEDGEMENTS

This article is based upon a presentation made at the Australian-American Joint Conference on the Technologies of Mines and Mine Countermeasures (MINWARA 99), July 12-16 1999, Sydney, Australia.

Thanks are due to Zheng Li for the GPR modelling and measurement results.

Lawrence Carter leads the Landmine Research Group at the University of Auckland, New Zealand. He is currently (April-September 2000) with the Joint Research Centre of the European Commission, Ispra, Italy.

http://www.esc.auckland.ac.nz/er/research/index.html
 
 

Last updated 9 Jannuary 2001 by GLR