Trial of Various Methods in Arid Soil at the Inshas Centre near Cairo

John F. CrawfordHengelweg 37, CH-5303 Wurenlingen, Switzerland

(Dated: May 16, 2008)

I. THE PROBLEM

As is well known, new ways of finding landmines areneeded. Among others, ground-penetrating radar (GPR)and neutron methods are under study. Both are affectedby moisture in the soil.

A. Ground-Penetrating Radar

Unlike pulsed radar, continuous-wave radars work bytransmitting at a certain frequency (typically for a mS)and then stepping to the next frequency. The reflectedsignal is measured in amplitude and phase at each fre-quency and stored. After the specified frequencies (typ-ically 256) have been scanned, a fast Fourier transformis carried out on a laptop. This converts from frequencyinformation to time (i.e. space) information, which isdisplayed. GPRs detect discontinuities in the dielectricconstant in the soil, such as a land mine.

In non-magnetic soil, RF energy penetrates a distanceset by the “skin depth”

δ =√

ρ× λ/Zo × π , (1)

where ρ is the soil resistivity, λ is the RF wavelength, andZo is the impedance of free space: about 120π Ω. Withincreasing soil moisture, both ρ and δ diminish, and theRF does not penetrate so well into the ground.

B. Neutron Methods

Explosives contain significant amounts of hydrogen; forexample, TNT contains 2.2% by weight of hydrogen, ormore meaningfully 24% by number. Therefore they mod-erate neutrons effectively. This effect can be exploited forexplosive detection as follows. Fast neutrons from a suit-able source such as Californium or Americium-Berylliumare moderated by the soil, and then detected by slow neu-tron counters, usually 3He proportional counters. Wherethere is an excess of hydrogen, e.g., in the explosive ina mine, there will be an enhanced count rate. Howeverthis effect depends on the soil not containing too muchhydrogen, the presence of which will prevent the neu-trons from penetrating and moderate them, weakeningthe signal. Thus in soil containing the same amount ofhydrogen as the explosive, e.g., as water, there will be nosignal at all.

FIG. 1: Photograph of the mines. They had been defused butwere otherwise real.

II. EGYPT AS A TEST AREA

Both GPR and neutron methods should perform bet-ter where the soil is dry, as is the case in North Africafor most of the year. Beyond that, magnetometry shouldbe effective against the steel-cased mines that are com-mon there. Given successful tests of these techniques inthe laboratory [1], a natural next step appeared to be toorganize a combined test in Egypt of as many devices as

TABLE I: Available Anti-Personnel (APMs) and Anti-TankMines (ATMs).

APM Quantity Charge (g) Total (g) Comments

VS-50 two 43 185 w/o 18 g

metal plate

PMN two 240 600

Box two ∼ 200 N/A one wood,

one plastic

ATM Quantity Charge (kg) Total (kg) Comments

T-80 two 4.5 N/A plastic

TM-46 one 5.7 8.6 steel

Israel one 10.5 N/A steel

M-71 one 6.25 9.8 steel

Mk-7 one 8.89 13.6 steel

2

FIG. 2: HYDAD-D (top) and PRIS (bottom) devices, withtheir operating crews. The string marking the test lane can beseen in both images. A minor problem was reading HYDAD–D’s oscilloscope screen in strong sunlight.

possible, especially GPR and neutronics methods. Theidea was first discussed at the IAEA TM-29225 meetingin Padova in November 2006, at which many of theselaboratory tests were reported.

A further advantage of running such a test in Egyptwas that it was possible to use real but defused land-mines, rather than surrogates. Lists of the anti-personneland anti-tank mines provided are shown in Table I. Theyare shown in Fig. 1.

Three stepped-frequency continuous-wave GHz GPRswere available. A team from Raumfahrt SystemtechnikAG, Salem, Germany (RST) brought two: HOPE andPRIS, both designed and built by RST [2]. HOPE worksfrom two to six GHz, and PRIS from 0.55 to 3.8 GHz.Beyond that, the Egyptian National Research Institute inAstronomy and Geophysics, (NRIAG) provided a GSSIMF20 GPR working at 0.1− 0.8 and 1− 2 GHz.

Two neutron detectors were available: HYDAD-D [3]from the University of Cape Town, South Africa; andESCALAD [4], built by a collaboration of Dutch and

Egyptian institutes.

III. METHOD

In a convenient part of the Inshas Centre of theEgyptian Atomic Energy Authority, two test lanes weremarked out in flat, dry, sandy soil by means of stringsstretched between pegs. Every 2 m, a position wasmarked with a knot. At each position an object wasburied, selected at random among mines, scrap metal,or nothing. Objects were covered to depths of 0, 10 or20 cm, again at random. Surface objects were coveredby just enough soil to conceal them. The soil at emptypositions was disturbed enough to avoid providing a vi-sual clue. The trials were “single blind” in the sense thatthe testees were not told which objects were in which po-sitions. Five APMs and five ATMs were buried in thetest lanes, the remaining two mines being required for si-multaneous tests with other equipment. HYDAD–D andPRIS, with their operators, are shown in Fig. 2.

IV. TEST LANE RESULTS

Results are shown in Tables II and III. NeitherHYDAD–D nor PRIS detects landmines as such; ratherthey detect proxies, in the form of hydrogen anoma-lies and radar reflections, respectively. Accordingly, thestatements “Positive” and “Negative” in the Tables re-fer to the apparent presence or absence of these proxies.“True” and “False”, on the other hand, refer to the pres-ence or absence of a land mine.

A. Notes on the APM test lane results

– The DLM2 calibrator was 4 m away from Position1; the other positions were 2 m apart.

– HYDAD–D took about 30 minutes to measure eachposition on the APM lane; this can be improved toonly a few minutes [5].

– PRIS took less than 60 minutes to measure alltwelve positions in a lane; these results are morefully described elsewhere [6].

– The MF20 GPR made a scan of the whole lane inonly a few minutes.

B. Notes on the ATM test lane results

– In a separate test, HYDAD–D saw such a strongsignal from some of these mines that part of thesoftware was confused.

– In separate test in a different area of the test field,PRIS missed one of the T-80 mines, apparently be-cause a cover plate was missing.

3

TABLE II: APM Lane Results. Column 3 shows the cover depth, column 6 the depth estimated by PRIS.

No. Object Depth (cm) HYDAD-D PRIS GPR Depth (cm) Remarks

0a DLM2 0 Calibration

1 Scrap 0 False Positive False Positive 5− 10

2 PMN 20 True Positive True Positive 15

3 Empty True Negative True Negative

4 PMN 10 True Positive True Positive 20

5 VS-50 20 True Positive False Negative See text

6 Scrap 20 True Negative False Positive 5− 10

7 Empty False Positive True Negative

8b Empty True Negative False Positive 15 See Note

9 Empty True Negative True Negative

10 Scrap 10 True Negative False Positive 15

11 Box 0 True Positive True Positive 5− 10 Wood

12c VS-50 10 False Negative True Positive 15 See Note

aDLM2 was a small dummy land mine supplied by the IAEA; itwas visible, and was used for calibration only.bHYDAD-D was misbehaving during this measurement; on its

return to Cape Town it turned out to have a fault that might havecaused it to miss this mine.cAlthough this position was nominally empty, three soft-drink

ring-pulls were buried here.

FIG. 3: PRIS Image of a PMN anti-personnel mine 8 cm deep.An estimate of the depth can be read from the vertical scale.

V. SOME EXTRA RESULTS

Although the trial concentrated on the test lanes, somefurther useful results were obtained:

1. As an exercise, a PMN was buried 15 cm deep. ThePRIS GPR was used to scan the area in the mannerof a metal detector. The PMN was found with nodifficulty.

2. A fluxgate magnetometer belonging to NRIAGcould detect steel ATMs at several metres. A

caesium magnetometer, some orders of magnitudemore sensitive, is on order.

TABLE III: ATM Lane Results. Column 3 shows the coverdepth, column 5 the depth estimated by PRIS. Because ofthe size of these mines, there were no false results and PRISwas able to estimate their depths fairly well. Time did notpermit HYDAD–D to be tested on this lane.

Number Object Depth (cm) PRIS Result Depth (cm)

1 M71 21 True Positive 15

2 T-80 21 True Positive 20

3 Empty True Negative

4 Empty True Negative

5 TM-46 35 True Positive 30

6 Mk-7 27 True Positive 20− 25

7 Empty True Negative

8 Empty True Negative

9 Empty True Negative

10 Israel 30 True Positive 25

11 Empty True Negative

12 Empty True Negative

4

FIG. 4: A image from the NRIAG fluxgate magnetometer.Mine 1 and Mine 2 show perturbations in the Earth’s mag-netic field of opposite sign. This presumably reflects theirindividual histories.

VI. COMMENTS & CONCLUSIONS

– We expected the dry Egyptian soil would help; itdid.

– A VS-50 - known to be a difficult mine for con-ventional metal detectors - was seen at a depth ofat least 20 cm by HYDAD–D. We believe this tobe a record. Given that the VS-50s provided werenot fitted with optional metal plates 1, this resultmay be competitive with what conventional metaldetectors would have been able to do.

– Because Egyptian soil is dryer than even a compa-rable sandy European soil, PRIS was able to seeabout 0.7 m into the ground – about twice as deepas in Europe.

– The UCT HYDAD–D device, and the RST GPRs,were transported as normal airline luggage - noteven excess baggage in the HYDAD–D case becausesome standard electronic units that it needs wereavailable in Cairo.

– Once minor problems had been overcome, theequipment was assembled and worked more or lessnominally. This indicates the state of developmentof the equipment.

– The weather during our test, 4-8 November, wasexcellent for our purposes, sunny, not hot by localstandards, and with light winds. Unusually, therehad been rain in October. If this had coincided

1 This mine can be supplied with an 18 g metal plate to make iteasier to find.

with our visit, some or all of our results might havebeen spoiled.

– ESCALAD gave a great deal of trouble, apparentlydue to interference from a nearby radio transmitter,to dust in electrical connectors, and to a mismatchbetween the strength of its neutron source and itsminimum practicable speed. ESCALAD is the firstdevice of its kind and this was its first field trialafter satisfactory laboratory tests [7].

Acknowledgments

This work relies heavily on the contributions of manypeople, in particular the following, whose contributionsit is a pleasure to acknowledge here.

Visiting Teams

– Ms Yvonne Krellmann and Mr Gunnar Triltzsch,RST, Salem, Germany: HOPE and PRIS GPRs

– Professor Frank Brooks, University of Cape Town,South Africa: HYDAD–D neutron detector

– Dr Victor Bom, Delft University of Technology,Netherlands, in collaboration with the EgyptianAEA: ESCALAD neutron detector

Local Participants

– From AEA, Cairo: Professor Riad Megahid and MrSalah Sendiony

– From the Inshas Nuclear Research Centre: ProfWagdy Ahmed Kansouh, Dr Ali Mostafa Ali,Messrs Ahmed Osman Abdo, Ashraf MostafaAbdel-Monem, Ibrahim Soliman Yousef, Mr SaadAhmed Atawy

– From NRIAG: Prof Magdy Atya, Prof Ibrahim El-Hemali, Dr Ashraf Khozayem, Dr Ahmed El Kotb,Dr Mamdouh Soliman.

– From the Tajoura Centre, Tripoli, Libya: DrFawzia El-Bakkoush and Ms Karima Shoshan

– From the Egyptian Army: General Ihab MohamedHelal

5

[1] IAEA Technical Meeting on Combined Devices forHumanitarian Demining and Explosives Detection,13-17 November 2006, INFN, Padova, Italy; seeSTI/Pub/1300 (2007) ISBN 978-92-0-157007-9.

[2] See www.rst-group.biz/index2.htm[3] F.D. Brooks and M. Drosg, Appl. Rad. and Isotopes 63

(2005) 565-74.

[4] V.R. Bom et al., accepted by Trans. Nucl. Sci.[5] F.D. Brooks, private communication[6] RST Report GPR-RST-RP-0022, available from RST,

Bahnhofstrasse 108, D-88682 Salem (Germany); E-mail:rst@w-4.de

[7] V.R. Bom, private communication