GPR and Metal Detector Portable System

Jean-Daniel Nicoud
LAMI-DeTeC, Swiss Federal Institute of Technology
IN-F Ecublens, CH-1015 Lausanne, Switzerland

Paper presented at the Symposium on Technology and the Mine Problem, Monterey, 18-20 Nov 96

Abstract

DeTeC (Demining Technology Center) is developing a sensor system for humanitarian demining, which reduces the number of false alarms and can be carried by a man or an autonomous lightweight robot.

The objective is to reliably recognize minimum metal antipersonnel mines. A metal detector is used to recognize the location of objects with some metal content. A GPR then provides an image that allows to differentiate a mine from metallic debris. The efficiency of deminers using the combined detector should increase significantly, and the database that can be built at the same time is essential for further steps in automating the search process.

Initially, deminers will look at the GPR images as an optional information, not changing their SOP (Standard Operation Procedures). They should progressively get confidence in the displayed information, which they can relate in real time with the result of their prodding.

1. Introduction

The metal detectors currently used by demining team cannot differentiate a mine from metallic debris, which sometimes leads to more than 1000 false alarms for every real mine found. Although the detectors can be tuned to be sensitive enough to detect the small amount of metal in modern mines, this is not practically feasible, as they will also be sensitive to ferrous soils, leading to the detection of smaller debris and augmenting the false alarms rate. Nowadays, once an alarm is given by the metal detector, the soil is prodded at a shallow angle using rigid sticks of metal to determine the shape of an object; this is an intrinsically dangerous operation.

The need for new, efficient and affordable demining technologies and sensor systems is therefore obvious. An overview of the current research status is given in [Maechler95] and [Gros96]. Past conferences dealing with this problem are listed in [Nicoud96b].

2. The DeTeC test system

Extensive tests in a "sand box" are required to develop the filtering and recognition algorithms, under repetitive conditions. Two containers have been built, one filled with sand and the other with loamy soil; they are 1 metre deep and 3.5 by 3.5 metres wide (Fig 1). A cartesian gantry positioning system allows to move the sensor above one of the containers at a time. Vertical motion is not controlled: the sensor is set at a fixed height, or a spring adjusts the pressure on the ground. The stepper motors control box receives displacement orders from a serial line, and the acquired data is stored on a PC's disk and transferred later to some server. Most of these measured data files are available on our Web site.

Figure 1: DeTeC test system: sand box, cartesian robot, 1 GHz radar antenna.

More realistic tests will be carried out at a later stage in the open. The cartesian positioning system is in fact easy to dismantle and carry. It just needs 4 support points for installation and can operate with the PC from a small power generator.

Tests are made with original inert mines and replicas, both very difficult to obtain. The explosive is replaced with wooden pieces of the same form, or explosive simulants such as beewax, or Dow Corning RTV 3110 and 3112 silicone rubbers [Bruschini96].

3. GPR selection

Current GPR systems are still way too expensive to be used in large number for humanitarian demining, such as it is now done with metal detectors. But we hope that prices will fall when the efficiency for mine detection will be proven and when the manufacturers will realize the potential market available.

A GPR for antipersonnel mine detection must have a wide frequency band to achieve a good resolution, but since higher frequencies do not propagate well, the chosen range is always a tradeoff between resolution and penetration depth. For antipersonnel mines (AP), a center frequency of 1 to 2 GHz, and a bandwith of the same magnitude, seem to be a good choice for most types of soil and for APs with a diameter of 8-10 cm. Smaller mines might require correspondingly shorter wavelengths, which will shorten the usable depth range too, but they are also buried closer to the surface.

A. Hardware

The radar choosen for our experiments is a SPRScan commercial system made by ERA Technology (UK). The acquired data is displayed in real time as a scrolling B-scan on the LCD screen of a rugged 486, 66 MHz PC. The antenna has a nominal bandwidth of 800 MHz to 2.5 GHz, which leads to an expected resolution of less than 5 cm.

All data are directly stored on the internal hard disk of the GPR and after that, files are transferred to a separate PC for data analysis. Most of them are freely available on Internet at http://diwww.epfl.ch/lami/detec/gprimages.html (SEG-2 file format used by the radar). Objects measured are antipersonnel mines and false positives (stones, bricks, wood and pieces of metal buried up to 30 cm). All these data are stored in one database and serve as input for algorithm evaluation.

B. Software

Software embedded in the radar is limited to some basic functions, mainly designed to improve the image quality and it is not sufficient for antipersonnel mine image analysis. Affordable GPR software for real-time applications seems not to be available on the market. Systems developed for military use are often mentioned, but are usually either classified or prototypes.

To start with, we have selected the Reflex seismic off-line processing package. Several modules are available for data analysis. Algorithms not included in Reflex are developed using the Matlab environment. Now, all the required modules are rewritten in Matlab, to allow us to evaluate all the modules of the data processing chain. The next step will be to rewrite all the routines for a fast DSP.

C. Data Visualization

Different visualization techniques are being evaluated to find the most suitable one, from a practical and computational point of view. One has also to bear in mind that in the demining case, GPR data will have ultimately to be interpreted by non expert personnel. The most common GPR data visualization consists in displaying the data as a vertical slice (Line or B-scan), whilst moving the antenna along a line on the surface.

If the real size of the buried target is needed by the recognition process, pulse deconvolution and migration algorithms will be necessary to transform the target response into a more compact one. We are still looking for a robust and fast algorithm which must be able to work on cluttered images. As soil characteristics play an important role in the migration aperture, it will also be useful to develop an adaptive algorithm.

In order to distinguish an object's shape it might be necessary to display horizontal views of the ground at different depths (Area or C-scan). In this case it is necessary to combine data from several parallel scans. The distance between two parallel scans is an important parameter, in order to reconstruct the real shape of the buried object. Parallel scans are performed each 20 mm, with an acquisition each 10 mm. In order to improve the resolution we take a second set of measurements orthogonally to the first one. The area of a minimum metal AP mine of diameter 8 cm is therefore covered by about 40 A-scans.

4. Induction coil sensor imaging

Instead of converting the information given by induction coil sensors to an audio signal, as it is done in conventional metal detectors, it is possible to use it for imaging purposes (displaying a map of the metal content in the soil), and to calculate a metallic object's parameters. With respect to this approach, the ODIS project at DASA-Dornier [Borgwardt95], in cooperation with the Foerster company, has demonstrated encouraging results.

The Foerster Minex 2000SL metal detector generates two continuous wave frequencies, f1 and f2, at 2.4 kHz (for ferromagnetic objects) and 19.2 kHz (for stainless steel and alloys) respectively. To fully exploit the detector's capabilities we intercept, at the output of the receiver-transmitter module, four signals corresponding (in the complex plane) to the real and imaginary parts of the analog signals f1 and f2 induced in the receiving coils.

5. Results

The response to the minimum metal mine (containing only a striker pin of 0.1 g!), a metallic debris of about 2 g and a stone (Fig 2) have been compared (Fig 3). Results are convincing, but at time of publication, the data acquisitions have been made in the sand box only, that is in a very clean environment.

Figure 2:The 3 objects used for initial comparative tests

Figure 3:GPR and MD images of the three objects

6. Hand-held device

Data acquisition in the sandbox benefits from the precise X-Y cartesian gantry. If the sensor is moved by hand, its position must be known precisely in order to rebuild an image comparable to the one extracted from the sand box test data. Irregular and redundant movements of the deminer must be sorted out and interpolated, in order to provide a x-y image. If the user is not correctly sweeping an area, he should be told to do some additional movements in a given direction.

Inertial sensors (2 accelerometers) are not acceptable, because even a slight inclination of the sensor head during the scan disturb the measure. We therefore choose to measure the distance with ultrasonic sensors (Fig 4). While the deminer is progressing within its security lane, two reflectors are moved along at each step. The area with precise enough measures (10mm) will be adjusted to 1.2m by 40cm.

Figure 4:Area covered by the distance sensor and typical sweep movement

The metal detector and GPR antenna cannot be superposed, since the Foerster differential metal detector has a coil in its center, which disturbs the GPR antenna. We had to juxtapose these elements. The total weight is important; an integrated design will be required before any production is started.

Figure 5:Ideas for packaging

7. Data fusion and identification

Before talking about fusing the GPR and metal detector image, an important database should be made available, acquired in a first step in the sand box, with real mines at different depths and orientations. A signature of the image should be extracted in order to reduce the data to be compared and fused.

In a first step, it is not required to identify precisely the mine. All mines must be signalled, with a safety better than 99.6%. They may then be prodded, or destroyed immediately. False alarms must be minimized, but a factor of 2, against the current 100 to 1000, of false alarms is probably acceptable.

Figure 6:Block diagram of the hand-held device

The files stored during operation (about 1 Gigabyte for one day's work) have two goals. First, on the same day, the demining supervisor will be able to visualize and comment to other deminers the decisions taken for some critical cases. Sharing experience will reduce the number of false alarms, hence increasing the efficiency of the team. Second, the accumulated database will allow to later train a neural network to take by itself the decision inside a future autonomous robot. A lightweight robot like the Pemex [Nicoud96b] has the potential to explore a complete field, mark the location of mines, and allows for their simultaneous destruction. Such a solution is for the moment too expensive to be acceptable by demining organizations.

Acknowledgments

This work is being supported by the Foundation "Pro Victimis" in Geneva, by the Swiss Department of Foreign Affairs and by the EPFL, who are cordially thanked.

All the work presented here has been performed by C. Bruschini, O. Carmona, B. Gros, F. Guerne, P-Y. Pièce, and M. Schreiber, who are congratulated for their engagement in this project and the very good results already obtained.

References

[Borgwardt95] C.Borgwardt, "ODIS - Ordnance Detection and Identification System", WAPM'95 Workshop on Antipersonnel Mine Detection and Removal, Lausanne, June 30-July 1st, 1995, pp 37-43

[Bruschini96] C.Bruschini et al., "Ground Penetrating Radar and Induction Coil Sensor Imaging for Antipersonnel Mine Detection" GPR'96, Sendai, Sept 30 - Oct 3, 1996, pp 211-216

[Garreau96] Ph.Garreau et al. "Potentials of Microwave to Tomographic Imaging for on-line Detection of Landmines", Detection of Abandoned Landmines, MD'96, Edinburgh, October 1996, pp164-166.

[Gros96] B.Gros, C.Bruschini, "Sensor technologies for the detection of AP mines: a survey of current research and system developments", ISMCR'96, Brussels, May 9-10, 1996, pp 564-569

[Fritzsche95] M.Fritzsche, "Detection of Buried Landmines using GPR and Metal Detector. First Results and Field Experiments", WAPM'95 Workshop on Antipersonnel Mine Detection and Removal, Lausanne, June 30-July 1st, 1995, pp 44-45

[Maechler95] Ph.Maechler, "Detection Technologies for Anti-Personnel Mines", Proceedings of the AS/MCM Autonomous Systems in Mine Countermeasures Symposium, Monterey, April 1995, pp.6.150-6.154

[Nicoud96a] J.D.Nicoud, Ph.Mächler, "Robots for Anti-Personnel Mine Search", Control Engineering Pratice, 4(4), April 1996, pp 493-498

[Nicoud96b] J.D.Nicoud, "Cooperation in Europe for Humanitarian demining", Symposium on Technology and the Mine Problem, Monterey, 18-20 Nov 1996 (this volume)

http://diwww.epfl.ch/lami/detec/monterey961.html | Last modified by CB 22.11.1996.

Go to DeTeC | EPFL