Home ATIP Publications ATIP Reports 2006 ATIP06.041: Demining Technology Development in Japan

 

ATIP in Action

Supercomputing Conference Series

ATIP participates in the Supercomputing (SC) conference series each year with a booth at the exhibition as well as our occasional workshops on HPC in Asia. Visitors to ATIP's booth can browse information, speak with analysts, request a guided tour of the Asian booths, and learn about our upcoming projects, activities, and events.  Please come visit us at SC14 in New Orleans, Louisiana , which is being held November 16-21, 2014.  We will be holding a workshop on "Japanese Research Toward Exascale Computing" on Monday, November 16, and will be presenting at booth #2920 all week during the Exhibition.

ATIP held a successful international panel discussion at SC13 featuring HPC experts from Australia, China, India, Japan, and Korea.  For more information, click here .  Copies of the presentations will be made available for download soon.

63

ATIP Panel at SC13. From left to right: ATIP's Dr. David KAHANER (Moderator), Prof. Depei QIAN (Beihang University, China), Dr. Kum Won CHO ( National Institute of Supercomputing & Networking, South Korea), Rob COOK (Queensland Cyber Infrastructure Foundation, Australia), Prof. N. BALAKRISHNAN (Indian Institute of Science, India), Prof. Taisuke BOKU (University of Tsukuba, Japan), and Panel Organizer Dr. Anne C. ELSTER, Director of the Heterogeneous and Parallel Computing Lab (HPC-Lab) at the Norwegian University of Science and Technology (NTNU)

ATIP's booth at SC12 in Salt Lake City, Utah

ATIP held a highly successful panel discussion called "Current Status of HPC Progress in China" at SC12 on Thursday, November 15, 2012. More than 100 conference attendees stopped in to check it out.  For more details on panel experts, click HERE.   

ATIP06.041: Demining Technology Development in Japan Print
User Rating: / 5
PoorBest 
ATIP Reports - 2006

ATIP06.041: Demining Technology Development in Japan

ABSTRACT: This report summarizes recent Japanese government-funded research on mine detection. The technologies being developed under a national project are based on a combined system utilizing metal detection (MD) and ground penetrating radar (GPR). In addition to these systems, direct explosive detection by neutron analysis and Nuclear Quadruple Resonance (NQR) are also being studied. An updated status of these technologies as well as an evaluation of their performance is provided within.

KEYWORDS: Environmental Remediation, Government Funding, Homeland Security, Imaging, Robotics, Sensors

COUNTRY: Japan


DATE:
27 September 2006

REPORT CONTENTS

  • INTRODUCTION EXECUTIVE SUMMARY
    • METAL DETECTOR AND GROUND PENETRATING RADAR SYSTEMS
    • 2.1 Overview
    • 2.2 Tohoku University’s SATO Laboratory
    • 2.2.1 System
    • 2.2.2 Performance and Evaluation
    • 2.3 The University of Electro Communication’s ARAI Laboratory
    • 2.3.1 System
    • 2.3.2 Performance and Evaluation
    • 2.4 Chiba University’s NONAMI Laboratory
    • 2.4.1 System
    • 2.4.2 Perfo rmance and Evaluation
    • 2.5 Development of a Mine Detection System at Nagoya University’s FUKUDA Laboratory
    • 2.5.1 System
    • 2.5.2 Performance and Evaluation
    • DIRECT EXPLOSIVES DETECTION TECHNOLOGIES
    • 3.1 Overview
    • 3.2 Neutron Analysis for Explosives Detection at Kyoto University’s YOSHIKAWA Laboratory
    • 3.2.1 System
    • 3.2.2 Performance and Evaluation
    • 3.3 Neutron Analysis for Explosives Detection at Nagoya University’s IGUCHI Laboratory
    • 3.3.1 System
    • 3.3.2 Performance and Evaluation
    • 3.4 Mine Detection by the NQR System at Osaka University’s ITOZAKI Laboratory
    • 3.4.1 System
    • 3.4.2 Performance and Evaluation
  • CONCLUSION
  • REFERENCES

1. INTRODUCTION

This report addresses demining technology developments in Japan. Research on this topic is current being conducted under a government project that focuses on humanitarian mine detection technology development. The Japan Science and Technology Agency (JST) manages the project, which was initiated in October 2002 and will end in 2007. Universities, companies, and government laboratories are cooperatively engaged in the research and development (R&D) for this project. The developed technologies have been practically evaluated in Croatia and Afghanistan.

Conventional mine detection is based on metal detectors (MDs). However, MDs have problems distinguishing between mines and metal fragments such as cartridge cases/shrapnel. False positives from MDs seriously reduce operational efficiency. Soils in Afghanistan and Croatia often contain large amounts of mineral matter, which also reduces the MD’s operational efficiency. In addition to these problems, a MD cannot detect plastictype mines.

In the JST project, combined systems utilizing MD and ground penetrating radar (GPR) are being developed as a short-term program. Application of GPR is expected to discriminate between mine and metal fragments by imaging buried materials. As a mid-term program, direct detection of explosive substances is being studied, utilizing techniques such as neutron analysis and Nuclear Quadruple Resonance (NQR).

ATIP offers a full range of information services, including reports, assessments, briefings, visits, sample procurements, workshops, cultural/business sensitivity training, and liaison activities, all performed by our on-the-ground multilingual experts. Email: This e-mail address is being protected from spambots. You need JavaScript enabled to view it Website: http://www.atip.org
Japan Office: ATIP Harks Roppongi Bldg, 1F 6-15-21 Roppongi Minato-ku Tokyo 106-0032 Japan Tel: 81 3 5411-6670 Fax: 81 3 5411-6671 U.S. Office (HQ): ATIP PO Box 4510 Albuquerque, NM 87196-4510 USA Tel: 1 (505) 842-9020 Fax: 1 (505) 766-5166 China Office: ATIP QingYun Modern Plaza, #2029 No. 43, W. Northern 3rd Ring Road Haidian District Beijing 100086 China Tel: 86 (10) 6213-6752 Fax: 86 (10) 6213-6732

Complete ATIP reports on Asian Science and Technology are provided to subscribers and collaborating organizations by direct distribution, or via electronic access. These contain text and often, charts, graphs, and pictures. Reports for unrestricted distribution often contain summarized or abstracted information. Sponsors can also obtain specific follow up information, including copies of proceedings, selected papers, exhibition particulars, updates, translations, query searches, etc.

END OF REPORT ATIP06.041a

ATIP06.041 (continued): Demining Technology Development in Japan

EXECUTIVE SUMMARY

  • Dual sensor systems utilizing metal detectors (MDs) and ground penetrating radar (GPR) are being developed in Japan for landmine detection; MDs work for small-size metal substance detection, and GPR discriminates between mines and debris/clutter by visually displaying images.
  • Through the introduction of GPR, dual sensor systems have been successful in detecting plastic landmines, which is impossible with systems that use only MD.
  • In terms of detectable depth, the performance of dual sensor systems is satisfactory; however, resolution is not sufficient.
  • Presently, professional deminers can outperform dual sensor systems in terms of detection speeds.
  • To directly detect explosive materials, neutron analysis and Nuclear Quadruple Resonance (NQR) are being developed as part of the national project.
  • Neutron generators are based on Deuterium-Deuterium (DD) reaction, but their sensitivity is not yet satisfactory. Improvement in the neutron generation rate is required for increased sensitivity.
  • A hand-held NQR system for explosive detection has been constructed, but the device’s sensitivity is not yet satisfactory. Increased sensitivity requires improvements in signal-tonoise ratio (S/N) reduction.

IMPACT & ASSESSMENT

In Japan, dual sensor systems, neutron analysis, and NQR are being developed for landmine detection. However, in terms of operational speed, sensitivity, and financial costs, these systems do not yet present advantages over operation by professional deminers. One potential area for practical application of the systems currently under development is the detection of plastic landmines, which cannot be detected by conventional MDs.

2. METAL DETECTOR AND GROUND PENETRATING RADAR SYSTEMS

2.1 Overview

Utilizing 10-KHz waves, MDs for mine detection can detect ~1 g of metal at a depth of 20 cm with an accuracy of 1 cm. However, a MD cannot detect plastic landmines, and it suffers from noise problems due to magnetized clutter in the soil. MD operators run the machine just above the ground’s surface by listening to delicate signal tones, which is a slow and inefficient process. The average scanable area in demining operations is reportedly an area of ~50m 2 /day.

Since GPR detection produces an on-screen image on a computer screen, the detection of plastic landmines is possible, in principle. Conventional GPR is designed for the inspection of water pipes or utility gas pipes buried ~1m below the ground. These pipes are usually more than 10 cm in diameter and are composed of either steel or poly-vinyl chloride (PVC). The pulse width of radar is set from a few to 10 nanoseconds (ns), which makes the inspection unit larger; therefore, the distance and direction resolution are not sufficient for mine detection. Mines are usually buried at depths from a few to 20 cm below the ground; mine detectors must be compact, and require innovations in both distance and direction resolution.

The MD and GPR system is a dual sensor system in which the MD works for small-size metal substance detection, while the GPR discriminates between mines and clutter through direct display imaging. Landmines are usually small in size, and explosives have electric properties similar to soil, which makes their radar reflection weak. In addition, random scattering caused by uneven ground surfaces interferes with the reflection from shallowdepth explosives.

Four universities in Japan -Tohoku, Electro-Communications, Chiba, and Nagoya -are currently working on a national landmine detection project utilizing a dual sensor system. A comparison of the characteristics and performance of each group’s system is presented in Table 1 below.

In terms of detection depth, three universities have reported operation depths of ~20 cm with their systems, while Chiba University has reported a depth of 12 cm. Considering a typical landmine’s buried depth, such performance appears to be satisfactory. All of the reported resolution values are within the range of 5–10 cm. Since landmines are usually small (around 1g), the resolution performance is still not sufficient for the detection of small landmines. However, the dual sensor system has demonstrated the possibility of plastic landmine detection to a depth of 20 cm, which is a major achievement of the project.

Operational time remains a major issue. The University of Electro-Communications (UEC) reports 200 sec/m 2 (18 m 2 /hr) and Nagoya University reports 300 sec/m 2 (12 m 2 /hr), respectively. According to Prof. FURUKAWA, the project’s counselor, a professional deminer probably sweeps ~50m 2 /day; the automated systems under development cannot match this value in field operations. In practical operation, a detection accuracy of 100% is required in order to avoid accidents. Reported values were obtained under conditions that cannot guarantee 100% accuracy, which could be the reason why reported values and field values are different.

Tohoku University’s system is small, but the other three systems are large and must be vehicle-mounted. Furthermore, operation over uneven ground is a common problem for all of the systems. Dr. Ikuo ARAI of UEC admits that financial cost will be an additional issue in the practical application of these systems.

Table 1. Comparison of Dual Sensor System Performance.

U Tohoku SATO UEC ARAI Chiba U NONAMI Nagoya U FUKUDA
System composition MD GPR MD GPR MD GPR Manipulator Robot MD GPR Manipulator
Weight & Size 25 kg (radar) 200 kg (manipulator)
Operation Sensor Technologies Hand-held Introduction of SAR-GPR to reduce clutter Vehicle-mounted High speed (3003,000 MHz, spiral antenna) GPR High resolution (150 ps pulse) Vehiclemounted Vehiclemounted 4 GHz GPR
Detection depth 20 cm (demonstrated) 20 cm (demonstrated) 12 cm (design) 10 cm clear 30 cm not clear
Detection speed NA 200 sec/m2 NA 0.5 m/sec (manipulator) 300 sec/m 2
Resolution >5 cm at 20 cm depth 1-2 cm (design) 10 cm at 10 cm depth
Plastic mines detectable detectable detectable detectable
Image display 2D & 3D 2D not clear 2D

2.2 Tohoku University’s SATO Laboratory

2.2.1 System

Prof. Motoyuki SATO’s team is developing two dual sensor systems for demining: The Advanced Landmine Imaging System (ALIS) and Synthetic Aperture Radar-GPR (SAR-GPR). The ALIS is a hand-held detector that can display objective images on the screen by receiving both MD and GPR signals. In the ALIS, the MD antenna and GPR coil are sophisticatedly allocated to prevent interference with one another. Information from MD, GPR, and sensor location is integrated by Kirchhoff migration algorithm. The operator scans and surveys the ground by looking at the screen. The sensor package of ALIS as well as MD and GPR images are shown below in Figure 1.

Prof. SATO’s team has also developed a SAR-GPR system (Figure 2). SAR can improve the radar’s horizontal accuracy with a small antenna. In Prof. SATO’s SAR-GPR system, moving antennas collect data. For moving antennas, clutter return signals are random, but signals from landmines are not. A newly developed algorithm analyzes the return signals and produces a noise-reduced signal from the landmine. The SAR-GPR is integrated with an antenna and vector network analyzer and is operated on the end of a manipulator.

2.2.2 Performance and Evaluation

Both the ALIS and SAR-GPR systems have been field-tested in Croatia and Afghanistan. SAR-GPR was able to detect landmines buried up to 20 cm in depth; MD alone could not detect landmines at this depth (see Figure 3). However, the SAR-GPR system recognized two landmines separated by 5 cm as one mine and failed to resolve them as separate objects. The ALIS could also detect both metal and plastic type mines at 20 cm depth. Images from the SAR-GPR look better than ALIS, due to the three-dimensional (3D) display. One advantage of Prof. SATO’s system is that it is hand-held and compact in size. Another advantage is that this system can adjust itself to soil moisture conditions, while

Figure 3. Images from the SAR-GPR.

conventional MD-GPR systems are less accurate under wet soil conditions. Thus, Prof. SATO’s system could be used at riversides, where robots cannot easily be used.

2.3 University of Electro-Communications’ ARAI Laboratory

2.3.1 System

Mine inspection must cover large areas. There is some expectation that GPR utilizing UHF band (300-3,000 MHz) radar can be utilized as high-speed mine detection radar. In Dr. Ikuo ARAI’s research, the application of narrow pulse width combined with an array of antennas has resulted in improvements in both resolution and operation speeds. For mine detection with an accuracy of 1-2 cm resolution in 2.0 dielectric soils, a pulse width (Tw) of 100-200 ps is required. Dr. ARAI’s unit uses a band pulse of 150 ps. Spiral antennas achieve broadband performance, and can be utilized without ground contact. Spiral antennas are normally small, but for soils with large attenuation, a larger attention is required since low frequency radio waves must be used. Dr. AKAI’s team fabricated a spiral antenna with a diameter of 9 cm. A number of spiral antennas are allocated in a flat configuration to operate at high speed (Figure 4).

Figure 5 below shows the Radar Unit mounted on a vehicle (manufactured by Fuji Heavy industries). The box at the front is the radar apparatus, which contains the Radar Unit, RF switch, switch controller, array antenna, and electric power source.

Figure 5. Radar apparatus mounted on a vehicle (left) and close-up of the vehicle-mounted Radar Unit (right).

2.3.2 Performance and Evaluation

Figure 6 shows mine images collected from depths of 10 cm and 20 cm, respectively. The image resolution of Dr. ARAI’s system is not sufficiently clear. The Radar Unit reportedly measures 75 cm x 24 cm x 36.5 cm and weighs 2 kg, but the system is vehicle-mounted rather than hand-held. The Radar Unit reportedly scanned mines buried underground at a speed of 1 cm/sec (100 sec/m), and completed scanning a 1 m 2 area within three minutes and 20 seconds (200 sec/m 2 ).

ATIP believes that two 1-meter-long scans will not be adequate for a 1 m 2 survey while maintaining a resolution of 1-2 cm. To maintain this resolution with 100% detection accuracy, more than two scans will be necessary. Dr. ARAI admits that the system needs further improvement in coping with uneven terrain, and he hopes to solve this problem with the application of multiple antennas. In addition, Dr. ARAI admits that the system is still too financially costly for practical application.

2.4 Chiba University’s NONAMI Laboratory

2.4.1 System

Dr. Kenzo NONAMI’s mine detection system is based on a robot equipped with a multifunction manipulator that has a GPR sensor and metal detector mounted on the end of the unit (see Figure 7 below). The multifunction manipulator is used to reveal mines buried near the surface, and can be equipped with a groundbreaking tool, a high-pressure air blower, and a gripper on the end of the manipulator. The gripper removes obstacles such as large stones and can handle obstacles weighing up to 10 kg. An electromagnet removes metal fragments such as shrapnel to improve the accuracy of mine detection.

Dr. NONAMI’s team’s prototype metal detector scans the ground, close to the surface. The metal detector consists of the metal detector body, two PC units, a 3D stereovision camera, and a two-dimensional (2D) camera. The detector has three degrees of freedom provided by a ball joint connected to three types of links. High-speed 3D mapping data is collected by a 3D stereovision camera, which is used to determine the best path for the metal detector to follow. The test range had uneven topography and contained plastic explosives buried at a depth of 12 cm. According to Dr. NONAMI, control of the sensor’s position is more important than attitude control.

Figure 7. The coordinate (articulation) system of the metal detector’s arm (left), and a view of the metal detector in the laboratory (right).

2.4.2 Performance and Evaluation

The manipulator reportedly weighs 200 kg, which is too large and too heavy for manual operation; thus, the system must be mounted on a vehicle. The designed performance for landmine detection is reportedly 12 cm in depth, and it targets plastic landmines. According to released pictures, the detector’s images of buried mines are less clear than those produced with other methods. The lack of clarity could be the result of a lack of fine control in positioning the metal detector. Dr. NONAMI admits that interference between MD and GPR is a serious problem; he does not currently have anything to report regarding practical operation results.

2.5 Nagoya University’s FUKUDA Laboratory

2.5.1 System

Dr. Toshiaki FUKUDA’s research team is developing a mine detection system that improves sensing performance, access, and control technology. The sensing unit is composed of an ultra-wide-band GPR with 4 GHz, coupled with a metal detector mounted on a terrain-hugging adjustable manipulator (Figure 8). The new unit improves position accuracy, and tracking along sloped ground while utilizing a 4GHz wave improves resolution. The manipulator can scan at a speed of 0.5 m/sec.

For demining, two types of vehicles are utilized (see Figure 9 below). An access vehicle parks near the area to be swept, and the sensing unit is carried on a boom mounted on the access vehicle. An assist

vehicle, from which an operator sends operation signals to the access vehicle, is parked behind the access vehicle. Due to the long boom arm, the system can scan a wide area; an eight-degree-of-freedom link mechanism has been developed to control the arm.

  • Performance and Evaluation
  • DIRECT EXPLOSIVES DETECTION TECHNOLOGIES

3.1 Overview

In dual sensor systems utilizing MD and GPR, detection is based on radar reflection and display imaging, but the systems do not provide information as to whether or not the detected matter is explosive. In 2004, JST initiated another program to directly check explosive materials utilizing neutron analysis and NQR. Mass spectroscopy is an important

11

method for detecting explosive materials, but in Japan, this system is used only in stationary applications such as ticket gate inspection, and not for landmine detection. For this reason, a discussion of mass spectroscopy technology is not included in this report.

Neutron analysis for landmine detection is based on the detection of specific gamma rays (γ rays) emitted when collisions between neutrons and Nitrogen (N) or Hydrogen (H) atoms occur. Among various γ rays, theγray at 10.8 MeV is crucial for the detection of N. For H detection, the γray at 2.22 MeV is used. The variance inγray radiation missions (depending on the N/H ratio in the explosives) allows for the determination of the type of explosives present. Key technologies are the neutron generator and the γray detector.

Although neutron analysis is used in oil and exploration, industrial inspection, and medical diagnosis, technologies for mine detection require great sensitivity, high-speed detection, and extremely high accuracy -all of which are more difficult to achieve than in conventional applications. The requirements for JST projects are as follows: 1) Greater than 200 mm detection for soil depth, 2) Less than 30g for minimum detectable explosive weights, 3) Detection accuracy greater than 50%, and 4) Total system weight less than 200 kg. A performance comparison of three direct explosive detection systems currently under development in Japan is provided in Table 2 on the following page.

Prof. Kiyoshi YOSHIKAWA’s group at Kyoto University is currently studying an inertial electrostatic confinement fusion neutron generator based on a Deuterium-Deuterium (DD) reaction. A Bi 4 Ge 3 O 12 (BGO) scintillator is utilized as the detector, and the application target is plastic-type mine detection. The diameter of the neutron generator is 30 cm, and the total weight of the system (including detector devices) is reportedly 150 kg. The system detects explosives and is operated by a 30-meter manipulator arm. The present neutron generation rate is 2.2x10 6 n/s, but Prof. YOSHIKAWA states that 1x10 8 n/s is required for practical applications. A demonstration test utilizing melamine (C 3 H 6 N 6 ) as a trigger explosive was conducted. When the distance between melamine and the detector was 12 cm, it took approximately 25 minutes to detect 60 g of melamine, which is far too slow for practical application at this time.

Prof. Tetsuo IGUCHI at Nagoya University also uses a DD generator for neutron analysis; the method is based on γray capture from thermal neutron and inelastic scattering γ rays emitted by fast neutrons. Prof. IGUCHI’s group is developing a multi-Compton camera-type detector based on a BGO scintillator as the γ ray detector, which can measure 10.8 MeVγrays and incidence direction simultaneously. Prof. IGUCHI developed an algorithm to obtain images of buried mines from a Compton gamma camera, which is a significant achievement of his research. Regarding sensitivity, the minimum weight of explosives required for detection is estimated as 30g at 5cm depth, and 200g at 20cm depth, respectively, with an accuracy of 50% and a detection ratio of 99.9 %. Prof. IGUCHI has made no statements about the size and weight of his system, but it will require vehicle-assisted operation.

Compared to Nuclear Magnetic Resonance (NMR), NQR does not require large magnets, and has the potential of being scaled down to a compact size. Additionally, this frequency band has a high ground penetration power that makes it a good tool for detecting objects located underground. The NQR resonance frequency of N atoms exists between 100 KHz and 10 MHz.

Professor Hideo ITOZAKI at Osaka University has successfully fabricated a hand-held NQR system that weighs less than 25kg and measures 550x250x400 mm. The electricity required at detection time is reportedly 45W.

For the demonstration of explosive signals, hexamethylenetetramine (HMT: C 6 H 12 N 4 ), a precursor of RDX (C 3 H 6 N 6 O 6 ), was used. With 25 g of HMT, the NQR signal is reportedly detectable at a distance of 10 cm, and with 500 g of HMT, the detectable distance is 30 cm. Sensitivity performance is not yet satisfactory, and an improvement in S/N reduction is required to increase the sensitivity.

Table 2. Performance Comparison – Direct Explosive Detection Technologies

System Neutron Analysis NQR
Researcher YOSHIKAWA, Kyoto IGUCHI, Nagoya ITOZAKI, Osaka
Neutron generator Inertial confinement (DD) Cockcroft accelerator (DD)
Neutron type Thermal Thermal & fast
Neutron generation 2.2x10 6 n/s at present 1x10 8 n/s target 1x10 8 n/s design
Detector BGO BGO Multi-Compton camera
Demonstration material Melamine (C3H6N6) Melamine
Detectable material weigh 30g @ 5cm depth 200g @ 20cm depth 25g @ 10cm 500g @ 30cm
Detection time 25 minutes @ 60g
System size or weight 150 kg Greater than 25 kg 550 x 250 x 400mm
Operation Vehicle-assisted Vehicle-assisted Hand-held
Accuracy 50% @ 99.9√ětection

The devices being developed by Profs. YOSHIKAWA and ITOZAKI meet JST’s target values for total system weight, while Prof. ITOZAKI’s NQR device has achieved hand-held operation. However, in terms of detectable explosive weight, the results are not yet satisfactory. Nagoya University’s system reports 50% accuracy, which is the JST target value.

3.2 Kyoto University’s YOSHIKAWA Laboratory

3.2.1 System

Prof. Kiyoshi YOSHIKAWA of Kyoto University is currently conducting research on an inertial electrostatic confinement fusion neutron source based on DD reaction. At the inertial electrostatic confinement fusion neutron source, Deuterium ions are generated by glow discharge, and these ions are accelerated at the center of a hollow cathode (Figure 11).

The system is classified as thermal neutron analysis. The application target is a mine detection device for plastic-type mines. Typical neutron generators utilize radioactive elements such as Tritium or Californium, which, according to the professor, are hazardous materials. His detector is based on monitoring H neutron scattering, which detects the presence of explosives as well as the variance in γray radiation emissions (depending on the N/H ratio in the explosives), which, as mentioned above, allows for a determination of the type of explosives present. The diameter of the neutron generator is 30 cm, and the total weight of the system (including detector devices) is reportedly 150 kg. The system is operated by a 30-meter manipulator arm. For the determination of N element, 10.83MeV γrays from N(n, γ) are measured by a particular detector system being researched by YOSHIKAWA’s colleague, Professor SHIROTANI, in which a BGO scintillator (Bi 4Ge3O12,

7.13 g/cm 3 ) is used. Background γ rays are rejected by an anti-coincidence method without the use of Lead (Pb), which reduces the overall weight of the system. To increase detection efficiency, escaped matter from the BGO detector is measured again by the surrounding Sodium Iodide (NaI) detector.

3.2.2 Performance and Evaluation

Thermal neutron analysis can detect N-containing explosives, but it cannot detect H-based and oxygen-based explosives such as triacetone triperoxide (TATP). The present neutron generation capability is 2.2x10 6 n/s, but Prof. YOSHIKAWA states that 1x10 8 n/s is required for practical application. Improving the device’s switching circuit could increase neutron generation.

The total weight of the system is reportedly 150 kg, which suggests the system is not suitable for portable use. Radioactive sources generally have the advantage of being financially less expensive and physically smaller than accelerator-type sources. Prof. YOSHIKAWA stated that Tritium, which is utilized in other types of neutron generators, presents a problem due to its radioactivity. However, the radiation produced by tritium is βrays, which can be shielded by a mere sheet of paper, suggesting that radioactivity might not be such a serious concern after all.

A demonstration test utilizing melamine (C 3 H 6 N 6 ) as a stimulant explosive was conducted. The amounts of melamine used were 800 g, 270 g, and 60 g, respectively. A container of melamine was placed at a distance of 30 cm from the neutron generator center, and the distance between the container and detector was 12 cm. The neutron generation rate was

2.2 x 10 6 n/s. In spite of the proximity between the system and melamine, and the large amount of melamine, it required about 1,500 seconds (25 minutes) to detect 60 g of melamine, which is far too long for practical application at this time. Prof. YOSHIKAWA said that increasing the neutron generation ratio by introducing a water-chilled generator and multiple detectors would shorten the time to a few minutes.

3.3 Nagoya University’s IGUCHI Laboratory

3.3.1System

Prof. Tetsuo IGUCHI’s method is based onγ ray capture from thermal neutrons and inelastic scattering of γ rays emitted by fast neutrons. His system is composed of a compact-size DD neutron generator and aγray detector.

For the neutron generator, Prof. IGUCHI applied a Cockcroft-type accelerator, and he used a helicon plasma ion source instead of a Penning-type ion source, due to the following reasons: 1) an expected, two-digit-higher neutron generation ratio; 2) pulse neutron generation; and 3) conjugated generation for monitoring. The schematic diagram of the generator is shown in Figure 12 below.

For mine detection, the neutron generator must have a neutron yield ratio greater than 2x10 8 n/s. To achieve this value, an acceleration voltage of 130 kV and an ion beam current of 3 mA peak value are required. In order to maintain constant neutron generation, it is necessary to keep Deuterium in a saturated condition at the titan target, which requires less than 150℃ at the target surface.

Prof. IGUCHI’s group is developing a multi-Compton camera type γray detector based on a BGO (Bi 4 Ge 3 O 12 ) scintillator (see Figure 13 below), which can measure 10.8 MeVγrays and incidence direction simultaneously. The detector is composed of six units that are equipped with 64 stacked scintillators and a Multi Anode Photomultiplier. The detector is designed to achieve a detection efficiency of nearly 40% for 10.8 MeV γrays, with a space resolution of ~10 cm at 25 cm.

Prof. IGUCHI developed an algorithm to obtain images of buried mines from the Compton gamma camera, which is a major achievement of his research.

3.3.2 Performance and evaluation

Figure 14 below shows an example of an image formed by Prof. IGUCHI’s system for the pseudo explosive, melamine.

In regards to sensitivity, the minimum weight of explosives required for detection is estimated to be 30 g at a depth of 5 cm and 200 g at a depth of 20 cm, with an accuracy of 50% and a detection ratio of 99.9%.

Prof. IGUCHI has made no comments regarding the system’s size and weight, but he displayed an image of a system that appeared to be sizeable (Figure 15). Based on the image, this system will require vehicle-assisted operation.

It should be noted that sensitivity is only an estimate and it has not been experimentally demonstrated. Prof. IGUCHI admits that further reduction of the S/N ratio is required to widen the detectable range and improve the accuracy.

3.4 Osaka University’s ITOZAKI Laboratory

3.4.1 System

Dr. Hideo ITOZAKI at Osaka University is studying the detection of explosives based on NQR. Since explosives typically contain nitrogen atoms, and the NQR resonance frequency of nitrogen atoms exists between 100 KHz and 10 Mhz, this process can detect the chemical structure of explosives through an induction coil or Superconducting Quantum Interference Device (SQUID). As mentioned earlier, NQR does not require large magnets as does NMR, and it has the potential of being scaled down to a compact size. Additionally, this frequency band has a high ground penetrating power that makes it a good tool for detecting objects located underground. However, it is difficult to narrowly focus this frequency band on small target areas due to its long wavelength, and it is also difficult to shield it from environmental noise. Figure 16 shows the NQR spectrums of some representative explosives.

THe SQUID built by Dr. ITOZAKI’s team can detect a wide range of magnetic field strengths (Figure 17). When a coil detector is used, sensitivity in the low frequencies is unsatisfactory, while the SQUID detector maintains good sensitivity throughout all frequency bands. Dr. ITOZAKI’s SQUID operates at the temperature of liquid nitrogen.

Dr. ITOZAKI’s system applies radio waves of 0.1MHz – 10MHz with 1 ms pulses. This system has the following advantages:

  • The waves can penetrate into deep soil and go around obstacles such as water.
  • The system can utilize technologies developed for radio, which are effective for a compact system.

Dr. ITOZAKI has succeeded in fabricating a hand-held NQR system, shown in Figure 18, but the system applies a coil detector rather than a SQUID. The system weighs less than 25 kg, and the reported size measure 550x250x400mm. The electricity required at detection time is 45W.

3.4.2 Performance and Evaluation

For the experimental detection of explosive signals in Prof. ITOZAKI’s demonstration, hexamethylenetetramine (HMT: C6H12N4), a precursor of RDX (C3H6N6O6), was used.

The amount of target material and distance between the material and detector are important factors for signal strength. With 25 g of HMT, the NQR signal is reportedly detectable at a distance of 10 cm; with 500 g of HMT, the detectable distance is 30 cm (Figure 19).

Although Prof. ITOZAKI has successfully fabricated a hand-held system, the following issues must be resolved:

  • Due to nuclear spin issues, this system can handle only N-containing explosives, and cannot detect non-nitrogen explosives such as TATP.
  • The signal from NQR is weak, so this system requires a high radio oscillator output of ~1 kW. In addition, the system requires short target distances and large amounts of explosives in order to obtain strong signals.
  • Shielding from oscillator and AM radio are important technologies for S/N ratio reduction.
  • Although SQUID can detect a wide range of magnetic field strengths, Dr. ITOZAKI admits that resolution at low-frequency bands, which corresponds to TNT analysis, is still not satisfactory. This could be due to technical difficulties with high-temperature super-conductive materials.
  • This system can detect explosives under moist soil conditions, but radar sensitivity is greatly reduced when the soil contains moisture.

4. CONCLUSION

Dual sensor systems utilizing MDs and GPRs are being developed in Japan for landmine detection; the MD works for small-size metal substance detection, and the GPR can differentiate between mines and non-mines by display imaging. Four universities in Japan – Tohoku, Electro-Communications, Chiba, and Nagoya – have joined a dual-sensor-system landmine detection project. By introducing GPR, dual sensor systems succeeded in the detection of plastic landmines, which MDs are not capable of detecting.

Regarding detection depth, three universities have reported results of ~20 cm with their systems. In view of practical buried landmine depths, this performance is satisfactory. All reported values are in the 5–10 cm resolution range; since landmines are usually small (~1 g), resolution performance is still not sufficient. In addition, operation time is another formidable problem. Professional “deminers” (i.e., humans) exceed dual systems in detection speeds. Additionally, although Tohoku University’s system is a hand-held type, the other three systems are large and require vehicle-assisted operation. Finally, operation on uneven ground is a common problem for all of the systems.

Dual sensor systems do not provide information as to whether or not the detected matter is explosive. In 2004, JST initiated another program to detect explosive materials directly, which includes neutron analysis and NQR.

Professor YOSHIKAWA at Kyoto University is currently studying an inertial electrostatic confinement fusion neutron generator based on a DD reaction. The system is operated by a 30-meter manipulator arm for plastic-type landmine detection. In a demonstration, it took ~1,500 seconds (25 minutes) to detect a 60-gram plastic explosive, which is far too long for practical application at this time. Present neutron generation capability is 2.2x10 6 n/s, and Prof. YOSHIKAWA states that 1x10 8 n/s is necessary for practical applications.

Prof. IGUCHI at Nagoya University also uses a DD generator for neutron analysis. He has developed an algorithm to provide an image of buried mines from a Compton gamma camera; this is a major achievement of his research. However, in terms of sensitivity, the performance is still not satisfactory.

Professor ITOZAKI at Osaka University has successfully fabricated a hand-held NQR system. Although sensitivity is close to the project target value, it is not sufficient. An improvement in S/N reduction is necessary to increase the level of sensitivity.

The systems under development all have potential application for the detection of plastic landmines, which conventional MDs cannot detect.

5. REFERENCES

ARAI, Ikuo Professor Electro-Communications University Dual Sensor

FUKUDA, Toshiaki Professor Nagoya University Dual Sensor

IGUCHI, Tetsuo Professor Nagoya University Neutron Analysis

ITOZAKI, Hideo Professor Osaka University NQR

NONAMI, Kenzo Professor Chiba University Dual Sensor

SATO, Motoyuki Professor Tohoku University Dual sensor

YOSHIKAWA, Kiyoshi Professor Kyoto University Neutron analysis

END OF REPORT ATIP06.041r