The work reported in this paper is a part of on-going studies to clarify how and to what extent soil electromagnetic properties affect the performance of induction metal detectors widely used in humanitarian demining. This paper studies the specific case of the time-domain response of a small metallic sphere buried in a non-conducting soil half-space with frequency-dependent complex magnetic susceptibility. The sphere is chosen as a simple prototype for the small metal parts in low-metal landmines, while soil with dispersive magnetic susceptibility is a good model for some soils that are known to adversely affect the performance of metal detectors. The included analysis and computations extend previous work which has been done mostly in the frequency domain. Approximate theoretical expressions for weakly magnetic soils are found to fit the experimental data very well, which allowed the estimation of soil model parameters, albeit in an ad hoc manner. Soil signal is found to exceed target signal (due to an aluminum sphere of radius 0.0127 m) in many cases, even for the weakly magnetic Cambodian laterite used in the experiments. How deep a buried target is detected depends on many other factors in addition to the relative strength of soil and target signals. A general statement cannot thus be made regarding detectability of a target in soil based on the presented results. However, computational results complemented with experimental data extend the understanding of the effect that soil has on metal detectors.

Most modern EMI mine detectors can detect the very small conductive and/or ferromagnetic parts of typical mines with relative ease. However, they also respond significantly to certain soils that contain lossy ferromagnetic minerals. In some special environments such as ocean beaches, conductivity of the host soil may also cause a response. Characterizing and modelling both the various target response mechanisms and the EMI detectors quantitatively would be relatively straightforward if it were not for the fact that most modern EMI detectors operate in time domain and use different current waveforms and time gates to observe response. Furthermore, much of the information about targets and interferences and even instrumental spectral limitations is observational rather than analytical data. In this paper, we put forward a spectral representation method that can be incorporated into both EMI data gathering and instrument modelling and which facilitates efficient quantitative simulation of arbitrary time- domain detection systems. The methodology and examples of its use are presented. Pure induction response from the ground is modelled with a sum-over-N-elements transfer function in which the kernel elements are single pole, pure damping responses which are log-spaced over the spectral range of interest. Instrument transfer functions can be described with a standard sparse pole and zero representation (located anywhere in the damped frequency half plane), if required. Model fitting techniques employing generalized inversion controls are used to go back and forth between frequency and time domain and the set of model parameters.

In the electromagnetic-induction (EMI) detection and discrimination of unexploded ordnance (UXO) it is important for inversion purposes to have an efficient forward model of the detector-target interaction. Here we revisit an attractively simple model for EMI response of a metallic object, namely a hypothetical anisotropic, infinitesimal magnetic dipole characterized by its magnetic polarizability tensor, and investigate the extent to which one can train a Support Vector Machine (SVM) to produce reliable gross characterization of objects based on the inferred tensor elements as discriminators. We obtain the frequency-dependent polarizability tensor elements for various object characteristics by using analytical solutions to the EMI equations. Then, using synthetic data and focusing on gross shape and especially size, we evaluate the classification success of different SVM formulations for different kinds of objects.

Conventional metal detectors are established and trusted tools for landmine detection, but their inability to precisely locate a target and discriminate mines from clutter leads to a high false alarm rate and slow rate of progress. This paper reports on developments to the Marmot advanced metal detector, which uses an array of coils to precisely locate a metal target in three dimensions and identify it. Recent developments allow the detector to calculate the magnetic polarizability tensor of a metal object. The magnetic polarizability tensor is unique to a particular target, and is a property of the metal's shape, size, conductivity, permeability and orientation. The eigenvalues of the magnetic polarizability tensor are compared to a library of values in the detector's software, representing common types of mine and clutter. In this way, Marmot can often quickly identify a detected object as a type of mine or a piece of clutter. This identification is independent of the target's orientation and, within limits, its position relative to the search head, thus providing the potential for a target recognition facility. This paper presents the results of tests to determine Marmot's ability to detect, precisely locate and identify common landmines. Tests have been conducted in air and in several types of soil. The instrument is a first step in developing the concept for landmine clearance. Issues for further investigation have been identified, including use of the instrument for identifying high metal content landmines, application of the soil rejection function and signal to noise issues.

The Johns Hopkins University Applied Physics Laboratory (APL) has developed a prototype metal detection survey system that will increase the search speed of conventional technology while maintaining high sensitivity. Higher search speeds will reduce the time to clear roads of landmines and improvised explosive devices (IED) and to locate unexploded ordnance (UXO) at Base Realignment and Closure (BRAC) sites, thus reducing remediation costs. The new survey sensor system is called the moving belt metal detector (MBMD) and operates by both increasing sensor speed over the ground while maintaining adequate sensor dwell time over the target for good signal-to-noise ratio (SNR) and reducing motion-induced sensor noise. The MBMD uses an array of metal detection sensors mounted on a flexible belt similar to a tank track. The belt motion is synchronized with the forward survey speed so individual sensor elements remain stationary relative to the ground. A single pulsed transmitter coil is configured to provide a uniform magnetic field along the length of the receivers in ground contact. Individual time-domain electromagnetic induction (EMI) receivers are designed to sense a single time-gate measurement of the total metal content. Each sensor module consists of a receiver coil, amplifier, digitizing electronics and a low power UHF wireless transmitter. This paper presents the survey system design concepts and metal detection data from various targets at several survey speeds. Although the laboratory prototype is designed to demonstrate metal detection survey speeds up to 10 m/s, higher speeds are achievable with a larger sensor array. In addition, the concept can be adapted to work with other sensor technologies not previously considered for moving platforms.

Electromagnetic Induction (EMI) is one of the most promising techniques for UXO discrimination. Target discrimination is usually formulated as an inverse problem typically requiring fast forward models for efficiency. The most successful and widely applied EMI forward model is the simple dipole model, which works well for simple objects when the observation points are not close to the target. For complicated cases, a single dipole is not sufficient and a number of dipoles (displaced dipoles) has been suggested. However, once more than one dipole is needed, it is difficult to infer a unique set of model parameters from measurement data, which is usually limited. Inspired by the displaced dipole model, we developed the dumbbell dipole model, which consists of a special combination of dipoles. We placed a center dipole and two anti-symmetric side dipoles on the target axis. The center dipole functions like the traditional single dipole model and the two side dipoles provide the non-symmetric response of the target. When the distance between dipoles is small, this model is essentially a dipole plus a quadrupole. The advantage of the dumbbell model is that the model parameters can be inferred more easily from measurement data. The center dipole represents the main response of the target, the side dipoles act as additional backup in case a simple dipole is not sufficient. Regularization terms are applied so that the dumbbell dipole model automatically reduces to the simple dipole model in degenerate cases. Preliminary test shows that the dumbbell model can fit the measurement data better than the simple dipole model, and the inferred model parameters are unique for a given UXO. This suggests that the model parameters can be used as a discriminator for UXO. In this paper the dumbbell dipole model is introduced and its performance is compared with that of both the simple dipole model and the displaced dipole model.

The generalized standardized excitation approach (GSEA) is presented to enhance UXO discrimination under realistic field conditions. The GSEA is a fast, numerical, forward model for representing an object's EMI responses over the entire frequency band from near DC to 100s of kHz. It has been developed and tested in both the frequency and time domains for actual UXOs placed in free space. The GSEA, which uses magnetic dipoles instead of magnetic charges as responding sources, is capable of taking into account the background medium surrounding an object. Given a modeled UWB frequency domain (FD) response, the corresponding time domain (TD) response is easily obtained by the inverse Fourier transform. Thus the technique is applicable for any FD or TD sensor configuration and can treat complex data sets: novel waveforms, multi-axis, vector, or tensor magnetic or electromagnetic induction data, or any combination of magnetic and EMI data. Host media effects are taken into account via appropriate types of Green's function and equivalent dipole sources. Comparisons between simulations and experimental data illustrate that the GSEA is a unified approach for reproducing both TD and FD EMI signals for actual UXOs. The EMI response from a soil that has a frequency-dependent magnetic susceptibility is studied. The EMI responses in both FD and TD domains are analyzed for the model of an actual UXO that is buried in a magnetically susceptible half space.

Electromagnetic induction (EMI) has been shown to be a promising technique for unexploded ordnance (UXO) detection and discrimination. The excitation and response of a UXO or any other object to EMI sensors can be described in terms of scalar spheroidal modes consisting of associated Legendre functions. The spheroidal response coefficients B^j_k correspond to the k^th spheroidal response to the j^th spheroidal excitation. The B^j_k have been shown to be unique properties of an object, in that objects producing different scattered fields must be characterized by different B^j_k. Therefore, the B^j_k coefficients may be useful in discrimination. We use these coefficients rather than dipole moments because they are part of a physically complete, rigorous model of the object's response. Prolate spheroidal coordinate systems recommend themselves because they conform most readily to the proportions of objects of interest. In clearing terrain contaminated by UXO, the ability to distinguish larger buried metallic objects from smaller ones is essential. Here, a Support Vector Machine (SVM) is trained to sort objects into different size classes, based on the B^j_k. The classified objects include homogeneous spheroids and composite metallic assemblages. Training a SVM requires many cases. Therefore, an analytical model is used to generate the necessary data. In simulation studies, the SVM is very successful in classifying independent sets of objects of the same type as the training set. Furthermore, we see that the B^j_k are not related to size or signal strength of the object in any simple or visually discernible way. However, SVM is still able to sort the objects correctly. Ultimately, the success of the SVM trained with synthetic (model derived) data will be evaluated in application to data from a limited population of real objects, including UXO.

There is a need for the stand-off detection of landmines either from a land or air based platform. Hyperspectral imaging technology has great potential for stand-off landmine detection. This paper will detail work undertaken by the UK Defence Science and Technology laboratory (Dstl) investigating the use of hyperspectral imaging for the detection of landmines. Both land and air based imagery has been collected using hyperspectral sensors in the VIS-SWIR region. This data and the initial results are discussed.

Illumination and polarization enhancement at optical wavelengths due to photometric backscattering by particulate surfaces can potentially be exploited for identifying soils disturbed by the burial of landmines, traversed by vehicles, and possibly for identifying environments prone to dust hazards. Laboratory and field measurements of reflected visible and near infrared light at phase angles from 3° to 0° demonstrate observations at these low angles are sensitive to grain size and packing for surfaces composed of particles as small as a few microns. Finer-grained materials tend to increase in reflectance more than do larger-grained surfaces and do so over a smaller range of phase angles, with a maximum brightness at or near 0° phase. Particulate surfaces also linearly polarize light; though the relationship is more complex, with high-albedo materials exhibiting a negative polarization of a few percent at one or few degrees phase. There is no intrinsic compositional dependency, and related albedo variations may be constrained using multiple wavelengths.

This paper describes the relative polarization and reflectance characterization of background and selected target items to demonstrate the differences material type and source wavelength have on these measurements. The advanced reflectance and polarization instrument (ARPI) was modified to allow three lasers with different wavelengths to be used. This allowed for similar spot size, location, and angles to be used to collect the measurements. ARPI was used to collect polarized and cross-polarized returns from the polarized laser source at an incident angle of 0, 5, 10, and 20 degrees. These measurements were used to calculate the relative percent polarization and percent reflectance. Analysis of the measured relative polarization and reflectance consists of single wavelength and multiwavelength comparisons with man-made and background items. A direct comparison is made between natural and man-made materials and different wavelengths of light. This careful comparison of differences between wavelengths will demonstrate which of the wavelengths produces the best and most consistent separation between background and manmade items. Our preliminary analysis shows that most man-made items give different polarization and reflectance returns than background items. Also, the analysis shows nominal variability between the three different wavelengths for background items and man-made items.

As a part of the Swedish mine detection project MOMS, an initial field trial was conducted at the Swedish EOD and Demining Centre (SWEDEC). The purpose was to collect data on surface-laid mines, UXO, submunitions, IED's, and background with a variety of optical sensors, for further use in the project. Three terrain types were covered: forest, gravel road, and an area which had recovered after total removal of all vegetation some years before. The sensors used in the field trial included UV, VIS, and NIR sensors as well as thermal, multi-spectral, and hyper-spectral sensors, 3-D laser radar and polarization sensors. Some of the sensors were mounted on an aerial work platform, while others were placed on tripods on the ground. This paper describes the field trial and the presents some initial results obtained from the subsequent analysis.

Traditional landmine detection techniques are both dangerous and time consuming. Landmines can be square, round, cylindrical, or bar shaped. The casing can be metal, plastic, or wood. These characteristics make landmine detection challenging. We have developed new methods that improve the performance of both surface and buried mine detection. Our system starts with the image segmentation based on a wavelet thresholding algorithm. In this method, we estimate the thresholding value in the wavelet domain and obtain the corresponding thresholding value in the image domain via inverse discrete wavelet transform. The thresholded image retains the pixels associated with mines together with background clutter. To determine which pixels represent the mines, we apply an adaptive self-organizing maps algorithm to cluster the thresholded image. Our surface mine classifiers are based on Fourier Descriptor and Moment Invariant to explore the geometric features of surface mines shown in the MWIR images. Our buried mine classifier utilizes the cluster intensity variations. To do this, we first cluster the target chip using a 3D unsupervised clustering algorithm. We then perform horizontal scanning to build a cluster intensity variation profile which is statistically compared with the signature profiles via Kolmogorov-Smirnov hypothesis test.

Creating a minefield requires disturbing the soil. This disturbance alters the soil properties and processes in a measurable way. The U.S. Army is investigating techniques to exploit the altered properties of disturbed soil to assist in the detection of buried landmines. The differential quartz reststrahlen signatures between disturbed and undisturbed soil at the long wave infrared (LWIR) region have shown promise in past field tests.^(1,3)We have initiated ground-based measurements using a non-imaging spectral sensor to investigate the phenomenology of LWIR disturbed soil signature. Our primary goal is to develop rainfall-dependent models to predict the degradation of the differential reststrahlen signature for varying soil types. A bare soil test site with strong quartz reststrahlen signature was selected for our initial investigation. The disturbed and undisturbed soil spectral signatures at the LWIR regions were obtained after multiple rain events using a Design and Prototypes field portable Fourier transform infrared (FTIR) spectrometer. The intensity and total amount of rainfall were recorded using a high-resolution tipping-bucket rain gauge. In addition to these measurements, photomicrographs of the disturbed soil were obtained after rainfall events, and X-ray diffraction analyses were conducted to obtain detailed soil mineralogy of the test site. We present these results and discuss the changes in the spectral characteristics of disturbed soil as a function of rainfall amount and intensity.

Research conducted at Georgia Tech over the past several years has focused on an examination of the signature characteristics of various background materials. These efforts seek to understand the physical basis and features of these signatures in order to aid the development of robust landmine detection techniques. The technical efforts in this study focused on identifying various soil types in LWIR hyperspectral imagery based on spectral characteristics. This paper will discuss the analytical approach and present results from these studies. A discussion of these terrain features as false alarm and clutter sources will also be presented.

Laser Induced Breakdown Spectroscopy (LIBS) is an emerging, minimally-destructive sensor technology for in-situ, real-time chemical species identification and analysis. The Army Research Laboratory has been engaged in LIBS analysis for over a decade and recently has been investigating the potential to apply broadband LIBS analysis to specific military problems, one of which is as a handheld, confirmatory sensor for landmine detection. Laboratory tests with a prototype man-portable LIBS system demonstrate a high degree of success in identifying landmine casings.

The objective of this paper is to present the Swedish land mine and UXO detection project named "Multi Optical Mine Detection System," MOMS. Research and investigations carried out within the MOMS project during the first year will be described. Activities have mainly been focused on basic principles, phenomena, acquisition of knowledge and literature studies. The paper introduces the reader with the aim of the project and then the initial and future work is presented.

Ferrimagnetic minerals such as magnetite and maghaemite can affect ground-penetrating radar (GPR) signals. This may lead to false alarms and missed targets when surveying for the detection of buried landmines and unexploded ordnance (UXO). In most field situations ferrimagnetic mineral content is too low to affect GPR wave behavior. However, in soils and sedimentary material with magnetite-rich parent material large concentrations of magnetite can be found. This paper is a first systematic experimental effort to study the effects of large concentrations of magnetite for GPR detection of subsurface targets. We study the effects of (i) different homogeneous mixtures of magnetite and quartz sand and (ii) magnetite concentrated in layers (placer deposits), on the propagation behavior of GPR waves and reflection characteristics of steel and plastic balls. The balls are buried in homogeneous mixtures of magnetite and quartz sand and below a layer of pure magnetite. Important observations include that the simulated placer deposits did have a large effect on the detectability of balls below the placer deposits and that homogeneous mixtures had no significant effect.

Information on the electromagnetic properties of soils and their effects on metal detectors is increasingly necessary for effective demining due to limited detector efficacy in highly magnetic soils and the difficulty of detecting minimummetal mines. Magnetic measurements of soils, such as magnetic susceptibility and frequency dependent susceptibility can aid the detection of problem soils, but are not part of standard soil analyses. Consequently, little information about soil magnetism exists within the soil, environmental science and environmental geophysics communities. Lack of empirical data may be compensated through the estimation of soil magnetic characteristics by predictive modeling approaches. Initial modeling of soil types in Bosnia and Herzegovina (BiH) was attempted by expert and analogue approaches, using only coarse scale soil type information, which resulted in the production of national soil maps for low field and frequency-dependent susceptibility. Validation of the maps was achieved by comparison of empirical magnetic data from soil samples in the National Bosnian soil archive in Sarajevo. Discrepancies between the model and empirical data are explained in part by the differences in soil parent material within each soil type, which controls the amount of Fe released into the soil system available for in situ conversion to magnetic Fe oxides. The integration of soil information (type and parent material), expert knowledge and empirical data refines the predictive modeling of soil magnetic characteristics in temperate-Mediterranean environments such as BiH. Further spatial separation of soil types in the landscape can be achieved by digital terrain modeling. Preliminary fine-scale, landscape-soil modeling indicates improved spatial resolution of soil types compared with the original coarsely-mapped soil units, and the potential to synthesize local scale soil magnetic maps.

Many soil physical and chemical properties interfere with landmine detection. Prior knowledge of these properties would improve detection technology selection and increase demining safety and efficiency. Developments in rapid mapping of these properties over wide areas is essential to meet military and economic constraints. Fusion of multiple detection technologies is also essential to overcome detection signal interferences. For these purposes, rapid mapping by use of remote sensing is being tested, starting with electrical conductivity mapping by radar remote sensing. Laboratory induced-polarization (IP) is also being tested to develop techniques to discriminate between electromagnetic signals from metallic particles in landmines and in soil, for regions with detection interference. Key physical models of soil are being developed for fusion of various landmine detection systems and to explain remote sensing responses to soil. Radar satellite tests carried out over the Canadian Forces Base Suffield (CFBS; Alberta, Canada) in 2004 and 2005 indicated 10 areas for possible high clay content and electrical conductivity. Eight of these were validated by soil maps and Landsat clay images. Two had high organic content with physical characteristics not known at present. Studies on soil with fine-grained iron-oxide powder and on iron with varied degrees of corrosion show that spectral-IP is sensitive to iron or iron-oxides regardless of their state. Soil has layered structure consisting of various grain-size combinations, but its physical characteristics are significantly influenced by whether its clay content is above or below a critical clay content (15 to 25 %). Results of these tests are discussed in this paper with explanations using the soil physical model.

In this paper we present the results of recent field and laboratory studies of the mineralogy and magnetic properties of young and/or weakly developed soils in Montana and California. The Chevallier Ranch UXO site in Montana is characterized by a basaltic plug and radiating feeder dikes, which is found surrounded by shales of the Spokane Formation. The site in California consists of an offset alluvial fan soil chronosequence of Little Rock Creek along the Mojave section of the San Andreas fault. The fan sediments include significant amounts of mafic material. The fan ages range from 16 to 413 thousand years. The results of magnetic susceptibility measurements and laboratory analysis of mineralogy demonstrate that the magnetic susceptibility in these soils is predominantly correlated with parent material and less with age or landscape position. Slow rates of soil forming processes lead to relatively low frequency dependence in magnetic susceptibility as compared to similar-age soils in tropical environments. The magnetic character of the soils can be accurately predicted with a previously developed model.

A variety of metal detectors are available for the detection of buried metallic targets in general and for humanitarian demining in particular. No one detector is optimal in all environments: variations in soil conductivity, and more importantly, frequency dependent soil magnetic susceptibility can favor one design over another. The use of computer modeling for assessing different designs is straightforward in principle, at least to first order, but still difficult in practice. The Geophysics Lab of the University of Toronto is attempting to address this problem in two ways. The first is by assembling the required computational algorithms to do this into a single simulation code with a straightforward GUI, intended to be public domain as a MATLAB code. The second, the subject of a companion paper in this conference, is by making measurements of the electromagnetic properties of difficult soils, and finding semi-analytic representations of these responses suitable for modeling purposes. The final version of the code, when completed, is to handle single or multiple transmitter and receiver coils of circular or polygonal shape, general transmitter current waveforms, arbitrary transmitter orientations and survey paths, small targets with frequency-dependent anisotropic responses (permitting both magnetic and inductive responses to be calculated), embedded in multi-layered half spaces with both conductivity and frequency-dependent susceptibility (so-called "difficult soils"). The current state of the simulation code and examples of its use will be described in this paper.

This paper analyzes the effect of the soil on the response of a metal detector (MD). The total response is first decomposed in a direct coupling between the transmitter and the receiver, the mine contribution and the soil contribution. The mine contribution is further related to its free space signature by introducting a number of transfer functions (TFs). Those TFs characterize the effect of the soil on the field propagation, from the transmit coil to the mine and back to the receiver, and on the mine signature. The expressions derived are quite general. However the TFs and other quantities of interest can only be computed if the scattering problem has been solved. For this it is usually necessary to resort to numerical techniques. Such techniques are computationally expensive, especially to analyze the various effects of the soil as they require to compute the solution for a large set of parameters. Therefore, we propose to model a buried mine by a multilayered sphere. From outside to inside, the layers represent the air, the soil, the mine explosive and the mine metallic content. Further, the analytic solution for such a multilayered sphere is used to compute the mine and soil responses, the mine free space signature and the various TFs as a function of the parameters of interest such as the soil electromagnetic (EM) properties or the mine depth. Finally, the validity domain of a number of practical approximations is discussed.

Magnetic and electromagnetic induction (EMI) sensing have been identified as two of most promising technologies for the detection and discrimination of subsurface metallic objects, particularly unexploded ordnances (UXO). In magnetic sensing, the principle of detection is that the sensor measures a distortion of the Earth's magnetic field caused by ferrous objects/ordnance. Similarly, in EMI, the sensors are detecting signals that are produced by induced and permanent magnetic polarizations. While these sensors can detect ferrous objects, they also find many other magnetic anomalies in the close vicinity. Soils, which contain small magnetic particles, called magnetically susceptible soils, can produce EMI responses, and therefore they can mask or modify the object's EMI response. These soils are a major source of false positives when searching for UXO using magnetic or EMI sensors. Studies show that in adverse areas up to 30% of identified electromagnetic (EM) anomalies are attributed to geology. Therefore, to enhance UXO detection as well as discrimination in geological environments the effects of the magnetic soils on the magnetic and EMI signal demands studies in detail. In this paper, the method of auxiliary sources (MAS) is applied to investigate the EMI response from magnetically susceptible rough surfaces. Several important physical phenomena such as the interaction between surface irregularities, modeled as multi hemitoroidal objects, surface roughness and antenna elevation effects are studied and documented. The numerical results are checked against available measurement data.

In this paper results are presented of a study on the performance of a dual-sensor landmine detector and its dependency on soil moisture. The detector was used on a trial site in the K5 mine belt in Cambodia. Soil samples were taken from the trial lane, as well as GPR measurements. The data obtained from these soil samples and field measurements are integrated into a model for soil moisture content that is correlated with the land mine detector performance.

In this paper, the characterization of the electromagnetic soil properties of a blind lane used in a trial for a dual-sensor mine detector is presented. Several techniques are used and are compared here; Time Domain Reflectometry, gravimetric techniques and Frequency Domain Reflection and Transmission methods. The derived soil properties are mapped by interpolation and the resulting maps are compared with the recorded deminers' performance on the lane. Recurrent and non-predicted results from the performance of the dual-sensor detector are explained as the results of variability of certain properties.

Magnetic soils confound both magnetometers and electromagnetic induction (EMI) sensors when these sensors are being used to detect landmines and unexploded ordnance (UXO). The amplitude of the magnetic susceptibility of the target is the problematic physical characteristic for magnetic detection; whereas the variation of magnetic susceptibility as a function of frequency, or magnetic viscosity, is the problematic physical characteristic that limits EMI sensor's effectiveness for target detection. Quantifying the physical characteristics of the soils in which targets are located can potentially provide insight into new methods of detection. Two new production sensors, which measure magnetic susceptibility as a function of frequency, have been tested on paramagnetic salts and soil samples from sites that exhibit magnetic viscosity. The purpose was to document their response for comparison with other popular sensors such as the Bartington MS2 system or the Quantum Designs MPMS. One of the new sensors, the SM-100 sensor from ZH Instruments, measures magnetic susceptibility at five fixed frequencies (~400Hz - 8 kHz) and six field strengths (10-320 A/m). The MAGNASAT sensor, a recently developed tool from Queensland Magnetic Research, can measure over a wider frequency band (10 Hz to 100 kHz) at a single field strength (80 A/m). The MAGNASAT measures the both the real and complex components of the magnetic susceptibility, whereas the SM-100 only measures the real component. Both sensors are sensitive enough to measure diamagnetic materials such as water, however, which is useful in field settings.

Modeling studies and experimental work have demonstrated that the dynamic behavior of soil physical properties has a significant effect on most sensors for the detection of buried land mines. An outdoor test site has been constructed allowing full control over soil water content and continuous monitoring of important soil properties and environmental conditions. Time domain reflectometry sensors and thermistors measure soil water¹ content and temperature, respectively, at different depths above and below the land mines as well as in homogeneous soil away from the land mines. During the two-year operation of the test-site, the soils have evolved to reflect real field soil conditions. This paper compares visual observations as well as ground-penetrating radar and thermal infrared measurements at this site taken immediately after construction in early 2004 with measurements from early 2006. The visual observations reveal that the 2006 soil surfaces exhibit a much higher spatial variability due to the development of mini-reliefs, "loose" and "connected" soil crusts, cracks in clay soils, and vegetation. Evidence is presented that the increased variability of soil surface characteristics leads to a higher natural spatial variability of soil surface temperatures and, thus, to a lower probability to detect landmines using thermal imagery. No evidence was found that the soil surface changes affect the GPR signatures of landmines under the soil conditions encountered in this study. The New Mexico Tech outdoor Landmine Detection Sensor Test Facility is easily accessible and anyone interested is welcome to use it for sensor testing.

The application of ground-penetrating radar (GPR) in humanitarian demining labors presents two major challenges: (1) the development of affordable and practical systems to detect metallic and non-metallic antipersonnel (AP) landmines under different conditions, and (2) the development of accurate soil characterization techniques to evaluate soil properties effects and determine the performance of these GPR-based systems. In this paper, we present a new integrated approach for characterizing electromagnetic (EM) properties of mine-affected soils and detecting landmines using a low cost hand-held vector network analyzer (VNA) connected to a highly directive antenna. Soil characterization is carried out using the radar-antenna-subsurface model of Lambot et al.¹ and full-wave inversion of the radar signal focused in the time domain on the surface reflection. This methodology is integrated to background subtraction (BS) and migration to enhance landmine detection. Numerical and laboratory experiments are performed to show the effect of the soil EM properties on the detectability of the landmines and how the proposed approach can ameliorate the GPR performance.

Buried in soil, landmines exhibit distinguishable nonlinear dynamic characteristics. These characteristics have been successfully used for nonlinear acoustic/seismic detection of both antipersonnel and antitank landmines. Despite a high potential of the nonlinear acoustic landmine detection technique, its utility is currently limited by a relatively high noise level of the LDV at frequencies typically used for landmine detection. To mitigate this limitation, we propose a modulation approach that exploits a nonlinear interaction of the low frequency resonance vibrations and higher frequency sound waves. The result of the modulation is manifested in a high frequency range as additional spectral components at the combination frequencies. The nonlinear response of the soil-mine dynamic system measured at the combination frequencies is used for the detection of the buried landmine. Exploring the higher frequency range has another benefit of using a directional high frequency sound source.

When airborne sound at two primary tones, f₁, f₂ (closely spaced near a resonance) excites the soil surface over a buried landmine, soil wave motion interacts with the landmine generating a scattered surface profile which can be measured over the "target." Profiles at the primaries f₁, f₂, and nonlinearly generated combination frequencies f₁-(f₂-f₁) and f₂+(f₂-f₁) , 2f₁-(f₂-f₁), f₁+f₂ and 2f₂+(f₂-f₁) (among others) have been measured for a VS 2.2 plastic, inert, anti-tank landmine, buried at 3.6 cm in sifted loess soil and in a gravel road bed. [M.S. Korman and J.M. Sabatier, J. Acoust. Soc. Am. 116, 3354-3369 (2004)]. It is observed that the "on target" to "off target" contrast ratio for the sum frequency component can be ~20 dB higher than for either primary. The vibration interaction between the top-plate interface of a buried plastic landmine and the soil above it appears to exhibit many characteristics of the mesoscopic/nanoscale nonlinear effects that are observed in geomaterials like sandstone. Near resonance, the bending (softening) of a family of increasing amplitude tuning curves, involving the vibration over the landmine, exhibits a linear relationship between the peak particle velocity and corresponding frequency. Tuning curve experiments are performed both on and off the mine in an effort to understand the nonlinearities in each case.

Acoustic methods of land mine detection rely on the vibrations of the top plate of the mine in response to sound. For granular soil (e.g., sand), it is expected that particle size will influence the mine response. This hypothesis is studied experimentally using a plate loaded with dry sand of various sizes from hundreds of microns to a few millimeters. For low values of sand mass, the plate resonance decreases and eventually reaches a minimum without particle size dependence. After the minimum, the frequency increase with additional mass includes a particle-size effect. Analytical continuum models for granular media applied to this problem do not accurately capture the particle-size effect. In addition, a continuum-based finite element model (FEM) of a two-layer plate is used with the sand layer replaced by an equivalent elastic layer. For a given thickness of the sand layer and corresponding experimental resonance, an inverse FEM problem is solved iteratively. The effective Young's modulus and bending stiffness of the equivalent elastic layer that match the experimental frequency are found for every layer thickness. Smaller particle sizes are shown to be more compliant in bending. The results clarify the importance of particle size on acoustic detection methods.

Recently, seismic landmine detection techniques have been investigated using ground-contacting sensors to measure ground motion generated by propagating surface waves and their interactions with buried objects such as landmines and clutter. Seismic waves have been generated using both ground-coupled and airborne sources, while non-contact sensors such as radar and laser-Doppler vibrometers have been preferred due to safety concerns. However, ground-contacting sensors can be effectively used provided that the contact with the ground does not adversely affect the propagation of seismic waves, that the sensor to ground coupling is repeatable, and that the sensors have low enough contact force to preclude triggering buried landmines. A groundcontacting sensor has been built with a low-cost commercially available accelerometer in a small lightweight package that ensures consistent coupling to the ground. Design and development of the sensor included experimental testing of several prototypes in a laboratory model, as well as analytical modeling of sensor response. A thirty-two-element line array capable of adjusting to surface contours of up to eight inches was tested at a U.S. Government facility in a temperate climate. The array enabled high-contrast detections of several AT landmines in both dirt and gravel roadbed sites.

Inexpensive ground-contacting accelerometers have been demonstrated in field experiments as appropriate vibrometers for a seismic landmine detection system. A thirty-two-element line array of these has been used to detect a variety of anti-tank (AT) landmines under realistic field conditions. Images of data measured by scanning this line array to synthesize a larger plane array have shown that the two-dimensional array offers potential advantages in terms of both measurement speed and landmine image contrast. The simultaneity of measurements with a physical array, as compared to synthetic array measurements that have been performed in the past, presents opportunities for improved landmine detection algorithms. Issues pertaining to the implementation of large arrays of vibrometers include sensor fidelity, array fidelity, scalability, and safety. Experimental measurements with prototype sensors in the laboratory and at a field test site have demonstrated the robust and repeatable ground coupling of the sensor in sand, dirt, gravel, and grass. Ground loading has been investigated with multiple array configurations with the dominant effect being an increase in the wave speeds of the surface waves. While the field experiments with the line array were conducted using commercially available data acquisition hardware and software, a custom data acquisition and processing system has been developed to meet the requirements of a large array of sensors. A lightweight sensor ensures the safety of touching the ground over buried landmines as the contact force is significantly less than the force required to detonate typical anti-personnel (AP) landmines and AT landmines.

In this paper, we consider the problem of detecting and locating buried land mines and subsurface objects by using seismic waves. We demonstrate an adaptive seismic system that maneuvers an array of receivers, according to an optimal positioning algorithm based on the theory of optimal experiments, to minimize the number of distinct measurements to localize the mine. The adaptive localization algorithm is tested using numerical model data as well as laboratory measurements performed in a facility at Georgia Tech. Cases with one and two targets are presented. It is envisioned that future systems should be able to incorporate this new method into portable mobile mine-location systems.

Several experiments have demonstrated the potential of Laser Doppler Vibrometry, in conjunction with acoustic-toseismic coupling or mechanical shakers, for the detection of buried landmines. For example, experiments conducted by The University Of Mississippi and MetroLaser, Inc. have shown the ability to scan a one square meter area in less than 20 seconds with a 16-beam multi-beam LDV (MB-LDV), and find the landmines under a variety of soil conditions. Some critical requirements for this technology are to reduce the measurement time, increase the spatial resolution, and reduce the size of the systems. In this paper, MetroLaser presents data from three optical systems that help achieve these requirements: 1) A Compact MB-LDV, 2) A two dimensional, or Matrix Laser Doppler Vibrometer (MX-LDV), and 3) A Whole-field Digital Vibrometer (WDV). The compact MB-LDV produces a 1-D array of beams, which may be scanned over the target surface with a scanning mirror. The size of the new, compact MB-LDV system has been reduced to approximately 17" x 11" x 9", thus enhancing its capability for field applications. The MX-LDV, to be developed in 2006, produces a 16x16 array of beams over a one meter area, allowing the ground velocity of the entire area to be measured in a single measurement. The WDV uses a camera-based interferometry system to take a snapshot of the ground vibration over a one meter square area with very high spatial resolution. Field tests for this system are scheduled for mid-2006.

The multi-beam laser Doppler vibrometer (MB-LDV) has been successfully used for acoustic landmine detection in field experiments at an Army test site. Using the MB-LDV in a continuously scanning mode significantly reduces the time of the measurement. However, continuous motion of a laser beam across the ground surface generates noise at the vibrometer output due to dynamic speckles. This speckle noise defines the noise floor and the probability of detection of the system. This paper studies the origins of speckle noise for a continuously scanning LDV. The structure of the speckle field exhibits points of phase singularity that normally coincide with signal dropouts. The signal dropouts and phase singularities can cause spikes in the demodulated velocity signal, which increase the noise in the velocity signal. The response of FM demodulators to input signals causing spikes in the LDV output are investigated in this paper. Methods of spike reduction in the LDV signals have been developed and experimentally investigated.

In this study, a new nonlinear acoustic technique, the phase shift method, is developed to measure the hysteretic nonlinearity parameter for a field soil. The technique is based on measuring the variation of phase difference between two transducers, i.e. the phase shift, induced by changing sound level. The hysteretic nonlinear parameter can be extracted from the measured phase shift as a function of sound level, or dynamic strain. With this technique, a long-term soil survey is conducted to study the variations of soil properties due to climate and seasonal changes. The hysteretic nonlinear parameter and sound speed of the soil as functions of temperature, moisture, surface tension, rain precipitation, and time are studied.

Probe force and ground surface velocity measurements are obtained using laser Doppler vibrometry for one specific excitation contact point on the casing of the VS1.6 Antitank landmine for surface laid and buried scenarios. Probe contact force and ground velocity measurements were taken over a 1 KHz bandwidth (0 Hz to 1 KHz). Combined velocity magnitude and phase images are provided as laboratory results. The proposed excitation technique has the potential for significantly greater signal bandwidth and amplitude compared to remote acoustic and seismic excitation strategies.

Time reversal acoustic (TRA) focusing allows concentration of elastic energy at a location in the soil being investigated to detect landmines. The TRA process is conducted by broadcasting a wide bandwidth signal and recording the surface vibration by a Laser Doppler Vibrometer (LDV). The system impulse response from speaker to the LDV output can then be computed by cross correlating the original and recorded signals for each channel. Each transducer re-radiates the time reversal impulse response. This provides efficient focusing of the seismic wave in both space and time, thus enhancing the nonlinear effects associated with soil and landmine vibrations. Using orthogonal initial signals the suggested TRA procedure can be implemented simultaneously with multiple transmitters to increase the scanning speed. The nonlinear effects were investigated using a phase inversion method where the TRA signal is broadcast a second time with an opposite sign and the two received signals are added in post processing. The summed signal contains mainly the results of nonlinear wave interaction and tends to cancel the linear response. Small scale land mine detection experiments were conducted using a laser Doppler vibrometer and an array of speakers in the frequency band 50-500Hz. They demonstrate that the TRA system provides high concentration of elastic wave energy in the tested area. The measurements of spectral density of the TRA focused signal reveal increased spectral density in the vicinity of mine resonance frequencies. The nonlinear TRA phase inversion method shows higher contrast between mine and no mine than the linear TRA method.

An improved sea mine computer-aided-detection/computer-aided- classification (CAD/CAC) processing string has been developed. The overall CAD/CAC processing string consists of pre-processing, subimage adaptive clutter filtering (SACF), normalization, detection, feature extraction, repeated application of optimal subset feature selection, feature orthogonalization and log-likelihood-ratio-test (LLRT) classification processing, and fusion processing blocks. The classified objects of 3 distinct processing strings are fused using the classification confidence values as features and either "M-out-of-N" or LLRT-based fusion rules. The utility of the overall processing strings and their fusion was demonstrated with new very shallow water high-resolution sonar imagery data. The processing string detection and classification parameters were tuned and the string classification performance was optimized, by appropriately selecting a subset of the original feature set. Two significant fusion algorithm improvements were made. First, a new nonlinear (Volterra) feature LLRT fusion algorithm was developed. Second, a repeated application of the subset Volterra feature selection/feature orthogonalization/LLRT fusion block was utilized. It was shown that this cascaded Volterra feature LLRT fusion of the CAD/CAC processing strings outperforms the "M-out- of-N," the baseline LLRT and single-stage Volterra feature LLRT fusion algorithms, and also yields an improvement over the best single CAD/CAC processing string, providing a significant reduction in the false alarm rate. Additionally, the robustness of cascade Volterra feature fusion was demonstrated, by showing that the algorithm yields similar performance with the training and test sets.

Over the past several years, Raytheon Company has adapted its Computer Aided Detection/Computer-Aided Classification (CAD/CAC) algorithm to process side-scan sonar imagery taken in both the Very Shallow Water (VSW) and Shallow Water (SW) operating environments. This paper describes the further adaptation of this CAD/CAC algorithm to process Synthetic Aperture Sonar (SAS) image data taken by an Autonomous Underwater Vehicle (AUV). The tuning of the CAD/CAC algorithm for the vehicle's sonar is described, the resulting classifier performance is presented, and the fusion of the classifier outputs with those of another CAD/CAC processor is evaluated. The fusion algorithm accepts the classification confidence levels and associated contact locations from the different CAD/CAC algorithms, clusters the contacts based on the distance between their locations, and then declares a valid target when a clustered contact passes a prescribed fusion criterion. Three different fusion criteria are evaluated: the first based on thresholding the sum of the confidence factors for the clustered contacts, the second based on simple binary combinations of the multiple CAD/CAC processor outputs, and the third based on the Fisher Discriminant. The resulting performance of the three fusion algorithms is compared, and the overall performance benefit of a significant reduction of false alarms at high correct classification probabilities is quantified.

Developing an effective detection and classification system for use with buried underwater objects is a challenging problem. In this paper, multichannel canonical correlation analysis (MCCA) is used for feature extraction from multiple sonar returns of buried underwater objects using data collected by the new generation Buried Object Scanning Sonar (BOSS) system. Comparisons are made between the classification results of features extracted by the proposed algorithm and those extracted by the two-channel canonical correlation analysis (CCA) algorithm on the SAX '04 data set. Extracted features are subsequently used in the development of classification systems able to differentiate between mine-like and non-mine-like objects. This study compares different feature extraction algorithms and classification schemes, and the results are presented in terms of classification rates and overall detection/classification performance. The results show that, for the SAX '04 data set, the features extracted via MCCA yield higher correct classification rates than feature extracted using CCA while simultaneously reducing structural complexity.

Many small Unmanned Underwater Vehicles (UUVs) currently utilize inexpensive, low resolution sonars that are either mechanically or electronically steered as their main sensors. These sonars do not provide high quality images and are quite dissimilar from the broad area search sonars that will most likely be the source of the localization data given to the UUV in a reacquisition scenario. Therefore, the acoustic data returned by the UUV in its attempt to reacquire the target will look quite different from the original wide area image. The problem then becomes how to determine that the UUV is looking at the same object. Our approach is to exploit the maneuverability of the UUV and currently unused information in the echoes returned from these Commercial-Off-The-Shelf (COTS) sonars in order to classify a presumptive target as an object of interest. The approach hinges on the ability of the UUV to maneuver around the target in order to insonify the target at different frequencies of insonification, ranges, and aspects. We show how this approach would allow the UUV to extract a feature set derived from the inversion of simple physics-based models. These models predict echo time-of-arrival and inversion of these models using the echo data allows effective classification based on estimated surface and bulk material properties. We have simulated UUV maneuvers by positioning targets at different ranges and aspects to the sonar and have then interrogated the target at different frequencies. The properties that have been extracted include longitudinal, and shear speeds of the bulk, as well as longitudinal speed, Rayleigh speed, and density of the surface. The material properties we have extracted using this approach match the tabulated material values within 8%. We also show that only a few material properties are required to effectively segregate many classes of materials.

This paper addresses the theories, experiments and real data of passive detection, classification and identification of "silent" targets in the illumination of ambient noise, a method known as "Acoustic Daylight." A great deal of work and sonar systems exist on active and passive sonar systems, but the principle of using ambient noise as the sole source of acoustic illumination was explored with limited success. This paper presents some of the successes using broadband signal processing and theory of target resonance as proposed in Uricks' text. In addition, the paper will present some of the results from experiments and simulations and Navy data of opportunities.

In this paper we present a time-frequency approach for acoustic seabed classification. Work reported is based on sonar data collected by the Volume Search Sonar (VSS), one of the five sonar systems in the AN/AQS-20. The Volume Search Sonar is a beamformed multibeam sonar system with 27 fore and 27 aft beams, covering almost the entire water volume (from above horizontal, through vertical, back to above horizontal). The processing of a data set of measurement in shallow water is performed using the Fractional Fourier Transform algorithm in order to determine the impulse response of the sediment. The Fractional Fourier transform requires finding the optimum order of the transform that can be estimated based on the properties of the transmitted signal. Singular Value Decomposition and statistical properties of the Wigner and Choi-Williams distributions of the bottom impulse response are employed as features which are, in turn, used for classification. The Wigner distribution can be thought of as a signal energy distribution in joint time-frequency domain. Results of our study show that the proposed technique allows for accurate sediment classification of seafloor bottom data. Experimental results are shown and suggestions for future work are provided.

A continuously adapting acoustic signal processor with an automatic detection/decision aid is presented. Its purpose is to preserve the signals of tactical interest, and filter out other signals and noise. It utilizes single sensor or beamformed spectral data and transforms the signal and noise phase angles into "aligned phase angles" (APA). The APA increase the phase temporal coherence of signals and leave the noise incoherent. Coherence thresholds are set, which are representative of the type of source "threat vehicle" and the geographic area or volume in which it is operating. These thresholds separate signals, based on the "quality" of their APA coherence. An example is presented in which signals from a submerged source in the ocean are preserved, while clutter signals from ships and noise are entirely eliminated. Furthermore, the "signals of interest" were identified by the processor's automatic detection aid. Similar performance is expected for air and ground vehicles. The processor's equations are formulated in such a manner that they can be tuned to eliminate noise and exploit signal, based on the "quality" of their APA temporal coherence. The mathematical formulation for this processor is presented, including the method by which the processor continuously self-adapts. Results show nearly complete elimination of noise, with only the selected category of signals remaining, and accompanying enhancements in spectral and spatial resolution. In most cases, the concept of signal-to-noise ratio looses significance, and "adaptive automated /decision aid" is more relevant.

Interactions of α-RDX, with the basal siloxane surface of the clay mineral kaolinite has been perform in our laboratory using the Gaussian 03 computational package. The low energy conformation of α-RDX (89.1 kcal/mol) was used to carry out the interaction with the clay mineral using the Hartree Fock (HF), DFT, DFT//HF, MP2, and MP2//HF levels of theory in order to determine the orientation, the types of bonds reacting between the two molecules and the adsorption as well. The results point out that the nitro group in pseudo-equatorial position interacts with the siloxane surface. DFT//HF level and Basis Set Superposition Error (BSSE) corrected MP2//HF were perform to obtain the binding energies (E_b) and the contribution of dispersion interaction to the binding energies (DE_b). The E_b using DFT//HF level of theory between those molecules fluctuate in a range of 38 to 51 kJ/mol once we applied the BSSE correction at different basis sets. Furthermore the results indicates a decrease in the E_b/BSSE (DFT//HF) of ~ 13 kJ/mol when polarization functions are added. The calculated binding energy of the RDX-siloxane surface complex is ~ 57 kJ/mol using MP2//HF/6-31+G (d) model chemistry. Studies of theoretical IR spectra of the interaction were obtained with DFT//HF methods and the 6-31+G(d) basis set with a small molecular model (single tetrahedra). The calculation predicted a band shift effect in the region of 1200-1800 cm^-1, due to interactions of the α-RDX with the siloxane surface.

A new type of optical chemical sensor recently developed in our lab has been demonstrated for highly sensitive, in-situ detection of explosives. The sensor is comprised of a dense silica thin film grown on the straight-cut endface of a standard, 125μm telecommunication optical fiber. Silicalite is an all-silica MFI-type zeolite with an effective pore size of 0.55nm. MFI zeolite is highly hydrophobic and selectively adsorbs organics of appropriate molecular size. The sensor device operates through measuring the optical refractive index or optical thickness of the coated zeolite film which changes in response to the adsorption of molecular species in its crystalline structure. In this work, the sensor exhibited different responses to simulants including pxylene, o-xylene, and triisopropylbenzene and trinitrotoluene (TNT) trace vapor in helium carrier gas.

New analytical methods have been developed and existing methods have been improved for the detection of explosives and their degradation products by increasing their sensitivity and selectivity. Some of the analytical methods available for detection of explosives and degradation products are gas chromatography, mass spectrometry, high performance liquid chromatography, and gas chromatography with mass spectrometry. This work presents the design and development of the experiments for the detection of the spectroscopic signature of TNT buried in sand and its degradation products. These experiments are conducted using a series of soil tanks with controlled environmental conditions such as: temperature, soil moisture content, relative humidity and radiation (UV and VIS). Gas chromatography and solid-liquid extraction with acetonitrile were used for the analysis of explosives. Sampling of tanks was performed in three points on the surface. The results show that TNT and 2,4-DNT are the main explosives that reach the surface of tanks. Temperature and water content play a most important role in the degradation and diffusion of TNT. Finally, the tanks were disassembled and sampling in deep with the objective to obtain a concentration profile. The results demonstrated that the highest concentration was located at 5 cm from surface.

We report on absorption-reflection measurements of ultraviolet and visible light from substrates contaminated with TNT. Bulk quantities of TNT are found to absorb ultraviolet-visible light between 250 and about 400 nm. No local maximum is found in the optical absorption spectrum of TNT, rather, the optical absorption decreases with wavelength to about 400 nm, with a high wavelength shoulder around 450 nm. Measurements at a distance are performed using the broad irradiation band of a 400 nm femtosecond laser pulse. Measurements performed at distances as large as 27 feet are consistent with TNT light absoprion around 400 nm.

PELAN (the initials of Pulsed Elemental Analysis with Neutrons) is a device that has been developed to identify a landmine through the elemental constituents of its explosive. PELAN uses neutrons as the probing particles. The incident neutrons interact with the nuclei of the various chemical elements in the mine, emitting characteristic gamma rays that act as the fingerprints of the various chemical elements. PELAN is capable of identifying with the same probability of detection all types of high explosives (TNT, RDX etc.) either in plastic or metal encased landmines. Results of its evaluation with blind tests in Croatia and the US using anti-personnel and anti-tank landmines will be presented.

In this paper we consider a linearised three-dimensional thermal model for buried landmine detection. Three aspects are analysed: i) efficient numerical methods for solving the thermal model, ii) estimation of physical and meteorological parameters of an arbitrary minefield which are actually necessary input parameters of the thermal model, iii) the validity and applicability of the thermal model in real minefields.

Identification and reduction of false alarms provide a critical component in the detection of landmines. Research at Georgia Tech over the past several years has focused on this problem through an examination of the signature characteristics of various background materials. These efforts seek to understand the physical basis and features of these signatures as an aid to the development of false target identification techniques. The investigation presented in this paper deal concentrated on the detection of foliage in long wave infrared imagery. Data collected by a hyperspectral long-wave infrared sensor provided the background signatures used in this study. These studies focused on an analysis of the statistical characteristics of both the intensity signature and derived emissivity data. Results from these studies indicate foliage signatures possess unique characteristics that can be exploited to enable detection of vegetation in LWIR images. This paper will present review of the approach and results of the statistical analysis.

Soil properties make a significant impact in the observed responses of various sensors for mine detection. Soil moisture affects the performance of electromagnetic sensors through its effects on soil thermal and dielectric properties. We have initiated laboratory, field and numerical studies to advance our fundamental understanding of the properties and governing processes of moisture distribution and flow around buried landmines. The laboratory component features magnetic resonance imaging (MRI) to map water distribution around a mine-like obstacle placed in a test soil sample. The field component investigates the moisture migration around landmines under realistic weather and soil conditions. We use anti-tank mines instrumented with moisture and temperature sensors to monitor the weather-driven processes. The numerical component investigates existing physics models underlying current simulations of moisture transport in soils. We use existing flow simulators to evaluate the completeness of process descriptions and to estimate the relative importance of individual processes on micro-scale moisture movement. These existing simulators include both continuum codes designed to work at scales much larger than the grain size and pore-scale models that discretize individual pores. We present the preliminary results of our investigations and discuss the potential impact of our findings on infrared and radar detection of buried landmines.

Soil moisture affects soil thermal and dielectric properties and may cause false alarms in detecting manmade objects when dielectric or thermal discontinuities exist in the soil. The spatial variability of soil moisture changes with time and it is important to understand this behavior because it is relevant for detection of small targets, and for modeling background moisture and temperature. Surface moisture of the top 6 cm of soil was sampled on regular grids with an impedance probe at a 0.1-m interval during wetting and drying events, both four days in duration. Maximum variances for data collected in August 2004 increased with decreasing mean moisture, as soil dried following a soaking rainfall. Maximum variances in June 2005 decreased over several days of intermittent rain as the soil rewetted following a prolonged drought. Spatially dependent ranges of approximately 0.5-m lag distance and exponential model fits were consistent among all the data sets, despite changes in moisture, moisture trend, and sample variance. The procession of spatial variation is described by variograms that transition from high to low maximum variances (sills) for wetting events, and from low to high maximum variances for drying events. A linear relationship between the maximum variance and mean of square root of ε was consistent for both years, except when the soil was incompletely wetted after a drought. The highest spatial variance in moisture that produced the most variable background for small target detection occurred as a consequence of the incomplete or uneven wetting following a drought.

The development of the scalable systems and methods involves proper reproduction of soil composition, lithology and structures, appropriate placement of boundary conditions, suitable simulation of representative environmental conditions, and the use of representative sampling systems. This paper evaluates the effect of different packing methods with a tropical sandy soil for obtaining a uniform and homogeneous packing so that these characteristics are comparable across all scales and dimensions. The packing methods used include piston-driven dry packing, piston-driven wet packing, and gravity-driven sedimentation packing. For dry and wet packing, the procedure consisted on the iterative addition of soil layers, mixing and compaction. Sedimentation packing involved the preparation of soil slurry and allowing its deposition under gravity. The systems were evaluated for consistent bulk density, porosity, homogeneity, and soil dispersivity. Preliminary results exhibit satisfactory bulk density and porosity values for the piston-driven methods, ranging from 1.59 to 1.64 g cm^-3 and from 42 to 44%, respectively. Sedimentation packing results in fair homogeneity and gradation, while dry packing develops heterogeneous layering. Transport parameters were also evaluated resulting in consistent dispersivity values for wet piston-driven packing ranging from 0.09-0.19 cm. Wet piston-driven packing is recommended as they yielded the most reproducible results for tropical sandy soils. The reproducibility of the recommended method is tested and proven in other physical models of different scales and dimensions. The method herein developed are, therefore, applicable for the development of representative multidimensional physical models designed to simulate soil and environmental fate and transport processes occurring in field conditions where landmines and other explosive devices are present.

Several military bases and monitions facilities throughout the world are contaminated with toxic explosives like 2,4,6-trinitrotoluene (TNT). This is an energetic compound and the least mobile of the military explosives. For this reason TNT gives one of the largest soil contamination problems. To understand the adsorption mechanism between TNT explosive and soil environments, the mechanical method analysis is used in our laboratory in order to obtain the soil texture classification. In these experiments, soil samples from horizons Ap and A were obtained from Jobos Series at Isabela, Puerto Rico. Based on the USDA texture triangle, the soil from the Ap horizon is classified as sandy clay. In contrast, the soil from A horizon fall in the sandy clay loam class. The clay minerals were separated from the other soil components using the mechanical method analysis. Cation exchange capacity (CEC), surface area, percent of organic matter and pH were determined for the soil and clay samples. The CEC results for soil samples were 3.62 mequiv/100 g for Ap horizon and 2.67 mequiv/100 g for A horizon while for the clay samples the CEC was 13.12 mequiv/100 g and 12.50 mequiv/100 g, for the Ap and A horizons, respectively. The results obtained for surface area analysis were 85.32 g/m² and 51.19 g/ m² for the two soil horizons and 189.71 g/m² and 163.87 g/m² for clay samples in the Ap and A horizons, respectively. These results indicate that the major adsorption could occur in the Ap horizon, specially in the clay fraction. A complete characterization of clay mineral samples using X-ray analysis reveals the present of kaolinite and quartz as main minerals. In order to obtain adsorption coefficients (K_d values), soil samples and the clay obtained from the mechanical method analysis, is being used for TNT adsorption studies by means of High Performance Liquid Chromatography (HPLC).

Deminers around the globe are still using handheld metal detectors that lack the capability to distinguish mines from clutter, detect mines containing very little metal, or find mines buried at deeper depths. In the southern African country of Angola, many areas and roads are impassable due to the threat of anti-tank landmines. Some of these mines are undetectable using current metal detector technology. The US Army has funded the development of the NIITEK ground penetrating radar (GPR) for detection of anti-tank (AT) landmines. This radar detects metal and plastic mines as well as mines that are buried too deep for handheld metal detectors to find. The US Department of Defense Humanitarian Demining (HD) Research & Development Program focuses on developing, testing, demonstrating, and validating new technology for immediate use in humanitarian demining operations around the globe. The HD team provided funding and guidance to NIITEK Incorporated for development of a prototype system called Mine Stalker - a relatively light-weight, remote-controlled vehicle outfitted with the NIITEK GPR, detection algorithms, and a marking system. Individuals from the HD team, NIITEK Inc, and the non-governmental organization Meschen Gegen Minen (MgM) participated in a field evaluation of the Mine Stalker in Angola. The primary aim was to evaluate the effectiveness and reliability of the NIITEK GPR under field conditions. The Mine Stalker was extremely reliable during the evaluation with no significant maintenance issues. All AT mines used to verify GPR performance were detected, even when buried to depths as deep as 25-33cm.

A complex of polarimetric (dual polarization), spatio-temporally combined active-passive devices of S (~3GHz), C (~5.6GHz), Ku (~20GHz), and Ka (~37GHz) band of frequencies is represented, for bare and vegetated soils, waved water surface and land snow cover microwave reflective and emissive characteristics simultaneous, multi-frequency, polarimetric and spatially coincident measurements. The complex is dedicated to solve problems applied to soil (bear and vegetated) and snow moistures retrieval, to near water surface wind and wave field parameters retrieval, by microwave means of remote sensing, as well as applied to surface and sub-surface targets detection and identification tasks solution. The complex is set in ECOSERV ROC's control-test experimental site, in Armenia, which is equipped by facilities for microwave devices absolute calibration, by spatially distributed stations for in-situ measurements of soil moisture and temperature, and has a local meaning small weather station. This paper has an aim to attract attention of researchers who are interested in such kind measurements and to invite them to perform their own or joint measurements using available facilities.

The assessment of the performances of ground-penetrating radar (GPR) in humanitarian demining is an important problem. These performances are related to the relative strength of the target radar response with respect to that of the soil. Many parameters influence both responses. The physical and geometrical parameters that influence the target signature include the soil electromagnetic (EM) constitutive parameters, the target depth and orientation with respect to the soil surface, the antenna height and the target EM and geometrical properties. This work presents a numerical parametric study of the soil and target radar signatures. The advantages of the numerical approach are: it allows for a separate study of the influence of each parameters on the radar responses, it is fast, cheap, generic with regards to hardware, and finally it is not prone to experimental errors and hardware failures or misuse. Moreover it is always possible to link the numerical experiments with a particular hardware by characterizing this latter. However, a number of simplifications, such as modeling the soil as a planar multilayered medium, are introduced to keep the problem tractable. This study yields surprising results, such as for example the possibility of considering the target in homogeneous space for computing its signature, as soon as it is a few centimeters deep. The target considered in the numerical experiments is a dielectric cylinder representing an AP mine, with diameter 6 cm and height 5 cm, and ε_rt=3. These values are chosen to approach as much as possible the physical properties of the M35BG AP mine, which is small and therefore difficult to detect.

Ground Penetrating Radar (GPR) has been applied for several years to the problem of detecting both antipersonnel and anti-tank landmines. RDECOM CERDEC NVESD is developing an airborne wideband GPR sensor for the detection of minefields including surface and buried mines. In this paper, we describe the as-built system, data and image processing techniques to generate imagery, and current issues with this type of radar. Further, we will display images from a recent field test.

A new study conducted using a Low Power Microwave Neutralization (LPMN) device is presented. The device consists of commercial off the shelf (COTS) magnetrons operating at 2.45 GHz and a simple power supply designed and fabricated at DRDC Ottawa that effectively doubles the average power of the magnetron. We describe the final design of a LPMN prototype device that was used in a live landmine field trial held at DRDC Suffield in the fall of 2004. Results of landmine trials using the LPMN to neutralize live munitions are shown. Using the LPMN to enhance landmine detection by infrared (IR) imaging is also described.

With over 110 million landmines buried throughout the world, the ability to detect and identify objects beneath the soil is crucial. The increased use of plastic landmines requires the detection technology to be able to locate both metallic and non-metallic targets. A novel active mmW scanning imaging system was developed for this purpose. It is a hyperspectral system that collects images at different mmW frequencies from 90-140 GHz using a vector network analyzer collecting backscattering mmW radiation from the buried sample. A multivariate statistical method, Principal Components Analysis, is applied to extract useful information from these images. This method is applied to images of different objects and experimental conditions.

Multi-static ground-penetrating radar (GPR) uses an array of antennas to conduct a number of bistatic operations simultaneously. The multi-static GPR is used to obtain more information on the target of interest using angular diversity. An entirely computer controlled, multi-static GPR consisting of a linear array of six resistively-loaded vee dipoles (RVDs), a network analyzer, and a microwave switch matrix was developed to investigate the potential of multi-static inversion algorithms. The performance of a multi-static inversion algorithm is evaluated for targets buried in clean sand, targets buried under the ground covered by rocks, and targets held above the ground (in the air) using styrofoam supports. A synthetic-aperture, multi-static, time-domain GPR imaging algorithm is extended from conventional mono-static back-projection techniques and used to process the data. Good results are obtained for the clean surface and air targets; however, for targets buried under rocks, only the deeply buried targets could be accurately detected and located.

SAR processing (or diffraction stacking migration) is an important signal processing method for ground penetrating radar (GPR) in many application case including landmine detection. It can improve signal-clutter ratio and reconstruct subsurface image, summing amplitudes along the hyperbolic trajectory that is Huygens surface (or diffraction travel time surface). For SAR processing, the travel time surface generally is smooth spherical surface in the case of zero-offset data and smooth ellipsoidal surface in the case of nonzero-offset data whose curvature is governed by the velocity function. But when the height of ground surface varies largely in the very rough ground area, for example mound, the travel time surface will be affected by the ground surface for the synthetic aperture - ground penetrating radar (SAR-GPR). In this paper we will consider the effect of the ground surface into the migration processing for the multi-offset CMP SAR-GPR data set. Firstly, using the SAR-GPR CMP data set, we will build the 3D velocity model including the ground surface topography and the subsurface velocity. Then, depending on the 3D model, we can do the ray tracing and compute the travel time between transmitter, receiver and each subsurface scattering point. At last, using the travel time, we can build the Huygens surface (or diffraction travel time surface) for each scattering point. The Huygens surface is the best migration trajectory. Depending on the Huygens surface and the migration trajectory of the SAR processing, we can discuss the migration aperture for SAR processing in rough ground area.

In this paper preliminary results of simultaneous and spatially coincident measurements of water surface, bare and vegetated soil and snow microwave reflective and emission characteristics at 37GHz is represented. The measurements were carried out by combined radar-radiometer system of Ka-band (~37GHz) of frequency set on a stationary platform of 6.5m of altitude.

This paper considers scattering by a buried, dielectric mine simulant at standoff distances of approximately 10 feet and depth one inch. It presents measured data, ray analyses, estimates of ray magnitudes for scattering centers, and an image formed from the data. The purpose was to understand the scattered field. Measurements, in the band 2 GHz to 6 GHz, used a fixed transmit antenna and a receive antenna that was moved over a linear path on one side of the transmit antenna. Complex reflectance was measured at one-inch intervals on the linear path. Range profiles were calculated by transforming reflectance over frequency for specific receive antenna positions. Separations of profile maxima agreed with ray path lengths fi-om the simulant's scattering centers. An image was calculated from two-dimensional spatial spectra produced by continuing the spatial spectrum of the measured data. The image had maxima near scattering centers, and it had another laterally displaced maximum. The displaced image maximum may result because the measured data maxima were spaced by 3 wavelengths, possibly generating a grating lobe. Further, reflectance was calculated as a sum of-rays-for the regions near reflectance maxima. Ray magnitudes from the front and back edges were generally larger than those for internal reflections, and magnitudes were largest near the center of antenna travel.

Standoff detection of landmines has long been central to improve quality of life in a number of countries around the world. A large body of work in the literature focuses on detection of TNT in soil as central to landmine detection. In this presentation, we summarize our efforts toward detection of TNT, from traces to bulk amounts, based on the absorption fingerprint of TNT. Light absorption by TNT is broken into three regions: (1) visible light absorption by TNT, and (2) formation and detection of NO₂ upon UV irradiation of TNT and (3) formation and detection of NO following UV absorption by NO₂. The absorption spectrum of TNT powder and particles has been determined from spectral analysis of backscattered visible light in traditional optical and near field optical microscopy measurements, respectively. The smallest amount of TNT detected in the near field measurements is 7 femtograms. The absorption spectra of TNT are rich in structure and similar to the one measured for gas phase NO₂, with lines due to roto vibronic coupling of electronic excited states. Measurements of the backscattered visible light on samples, placed about 5 to 10 meters from the laser source, indicate a clear change in intensity as compared to samples containing TNT. Turning to the second light absorption region, NO₂ is detected upon UV irradiation of solid TNT. NO can also be detected by photolysis of NO₂.

We use femtosecond laser mass spectrometry (FLMS) to study the fragmentation patterns of solid phase explosive materials subjected to femtosecond laser pulse irradiation. In condensed phase FLMS a compound deposited on a solid substrate is desorbed into vacuum by femtosecond irradiation forming a plume of ionized and neutral species. Positive or negative ions are accelerated by an electric potential, allowed to drift in the field-free region of a time-of-flight (TOF) mass spectrometry instrument, and flight-times are recorded by a micro-channel plate detector and a digital oscilloscope. From the value of the accelerating field and the ion flight time, the mass-to-charge ratio of each ion is obtained. In this paper we report femtosecond laser mass spectra for the positive and negative ions formed by desorbing TNT and RDX with 150 fs pulses centered at 800 nm. The fragmentation pathways for the formation of the observed ions are described and are used to interpret femtosecond laser induced breakdown spectroscopy results.

We use femtosecond laser-induced breakdown spectroscopy (LIBS) to detect trace amounts of TNT and RDX. A high-power pulsed laser is used in LIBS to form a plasma on the material surface and the optical emission from the plasma is spectrally analyzed to determine the material composition. Femtosecond LIBS results for TNT and RDX on aluminum substrates and glass slides are reported. Results are examined in terms of the optical properties of the substrate and the strong linear absorption for aluminum is contrasted with the weaker multiphoton absorption for glass. Optical microscope images of the ablated explosives are shown for femtosecond and nanosecond laser excitation. Fragmentation studies by femtosecond laser mass spectrometry are used to interpret LIBS results.

Genetically modified plants that detect TNT and its degradation products are potentially powerful aids in humanitarian demining and detection of unexploded ordnance. Although the feasibility of TNT detection by plants and microorganisms has been demonstrated by several research teams world wide, thus far, none of these previously demonstrated systems has the sensitivity and specificity to be effective under field conditions. We are using two approaches to increase the potential effectiveness of these and related biological detection systems. First, we are expanding the repertoire of explosive-responsive promoters by conducting DNA microarray experiments with plants treated with TNT-degradation products, and characterizing the inducibility of reporter gene expression by these promoters. Second, we are evaluating the dynamics and limiting factors in the transmission of artificial signals from roots to shoots. This will increase the ability of soil-based TNT perception strategies to effect human-readable changes in shoot morphology as part of a practical plant-based explosives detection system.

Nuclear methods have long been one of the few techniques available to aid in the detection and identification of potentially dangerous objects in a non-intrusive manner. The application of neutron-based methods has been particularly successful in bulk material detection and identification, owing to the neutron's capability to penetrate deep into materials, and its nuclide-specific interactions which can be used to make direct measurements of a target's elemental composition. Defence R&D Canada - Suffield's initial work in the area of penetrating radiation resulted in the development of the recently commercialized Minespec, a Thermal Neutron Analysis (TNA) system for buried-explosives detection. Co-developed with Bubble Technology Industries Inc., as the confirmation detector for a multi-sensor anti-tank landmine detection system, continuing improvements to the TNA system have included the inclusion of an electronic pulsed neutron generator - an upgrade that presents the possibility of utilizing Fast Neutron Analysis (FNA) methods to improve the system's detection capability. In this paper we will discuss the Minespec system and report on our investigations regarding the possibility for incorporating an FNA component to provide complementary information to assist in anti-tank landmine detection.

Nuclear Quadrupole Resonance (NQR) is a radio frequency (RF) technique that can be used to detect the presence of quadrupolar nuclei, such as the ¹⁴N nucleus prevalent in many explosives and narcotics. The technique has been hampered by low signal-to-noise ratios and is further aggravated by the presence of RF interference (RFI). To ensure accurate detection, proposed detectors should exploit the rich form of the NQR signal. Furthermore, the detectors should also be robust to any remaining residual interference, left after suitable RFI mitigation has been employed. In this paper, we propose a new NQR data model, particularly for the realistic case where multiple pulse sequences are used to generate trains of spin echoes. Furthermore, we refine two recently proposed approximative maximum likelihood (AML) detectors, enabling the algorithm to optimally exploit the data model of the entire echo train and also incorporate knowledge of the temperature dependent spin-echo decay time. The AML-based detectors ensure accurate detection and robustness against residual RFI, even when the temperature of the sample is not precisely known, by exploiting the dependencies of the NQR resonant lines on temperature. Further robustness against residual interference is gained as the proposed detector is frequency selective; exploiting only those regions of the spectrum where the NQR signal is expected. Extensive numerical evaluations based on both simulated and measured NQR data indicate that the proposed Frequency selective Echo Train AML (FETAML) detector offers a significant improvement as compared to other existing detectors.

Sensor fusion has become a vital research area for mine detection because of the countermine community's conclusion that no single sensor is capable of detecting mines at the necessary detection and false alarm rates over a wide variety of operating conditions. The U. S. Army Night Vision and Electronic Sensors Directorate (NVESD) evaluates sensors and algorithms for use in a multi-sensor multi-platform airborne detection modality. A large dataset of hyperspectral and radar imagery exists from the four major data collections performed at U. S. Army temperate and arid testing facilities in Autumn 2002, Spring 2003, Summer 2004, and Summer 2005. There are a number of algorithm developers working on single-sensor algorithms in order to optimize feature and classifier selection for that sensor type. However, a given sensor/algorithm system has an absolute limitation based on the physical phenomena that system is capable of sensing. Therefore, we perform decision-level fusion of the outputs from single-channel algorithms and we choose to combine systems whose information is complementary across operating conditions. That way, the final fused system will be robust to a variety of conditions, which is a critical property of a countermine detection system. In this paper, we present the analysis of fusion algorithms on data from a sensor suite consisting of high frequency radar imagery combined with hyperspectral long-wave infrared sensor imagery. The main type of fusion being considered is Choquet integral fusion. We evaluate performance achieved using the Choquet integral method for sensor fusion versus Boolean and soft "and," "or," mean, or majority voting.

The Improved Landmine Detection System (ILDS) is intended to meet Canadian military mine clearance requirements in rear area combat situations and peacekeeping on roads and tracks. The system consists of two teleoperated vehicles and a command vehicle. The teleoperated protection vehicle precedes, clearing antipersonnel mines and magnetic and tilt rod-fuzed antitank mines. It consists of an armoured personnel carrier with a forward looking infrared imager, a finger plow or roller and a magnetic signature duplicator. The teleoperated detection vehicle follows to detect antitank mines. The purpose-built vehicle carries forward looking infrared and visible imagers, a 3 m wide, down-looking sensitive electromagnetic induction detector array and a 3 m wide down-looking ground probing radar, which scan the ground in front of the vehicle. Sensor information is combined using navigation sensors and custom navigation, registration, spatial correspondence and data fusion algorithms. Suspicious targets are then confirmed by a thermal neutron activation detector. The prototype, designed and built by Defence R&D Canada, was completed in October 1997. General Dynamics Canada delivered four production units, based on the prototype concept and technologies, to the Canadian Forces (CF) in 2002. ILDS was deployed in Afghanistan in 2003, making the system the first militarily fielded, teleoperated, multi-sensor vehicle-mounted mine detector and the first with a fielded confirmation sensor. Performance of the prototype in Canadian and independent US trials is summarized and recent results from the production version of the confirmation sensor are discussed. CF operations with ILDS in Afghanistan are described.

SAR-GPR is a sensor system composed of a GPR and a metal detector for landmine detection. The GPR employs an array antenna for advanced signal processing for better subsurface imaging. This system combined with synthetic aperture radar algorithm, can suppress clutter and can image buried objects in strongly inhomogeneous material. SAR-GPR is a stepped frequency radar system, whose RF component is a newly developed compact vector network analyzers. The size of the system is 30cm x 30cm x 30cm, composed from 6 Vivaldi antennas and 3 vector network analyzers. The weight of the system is less than 30kg, and it can be mounted on a robotic arm on a small unmanned vehicle. In the signal processing of the SAR-GPR has a unique future. It can be used with an algorithm for strong clutter suppression. The sensor has about 10cm offset from the ground surface, and it can even image the ground surface topography. It will be implemented for more advanced imaging algorithm, which can be used for the ground surface with a large roughness. Field tests of SAR-GPR were carried out in March 2005 in Japan. Then after, it was also evaluated in the Netherlands and Croatia. We report the results of these evaluation and demonstration.

The Autonomous Mine Detection Sensors (AMDS) program is developing a prototype autonomous mine-detection sensor suite designed to be mounted on a small robotic platform that can find buried anti-personnel mines. Over the past two years, CyTerra Corp. and NIITEK, Inc. have developed complementary senor suites using a variety of ground penetrating radar (GPR) and electromagnetic induction (EMI) sensor configurations. The AMDS program is also working with industry and academia to develop automatic target recognition (ATR) algorithms. This paper provides a brief overview of evaluations that have been performed at Army facilities. Probability of Detection (Pd) and Probability of False Alarm (Pfa) results are provided for signal-to-noise type detection algorithms and also for promising pattern classification and neural network algorithms that were developed by Duke University, the University of Missouri-Columbia, and the University of Florida. After an evaluation in October 2005, both contractors' sensors performed comparably (about 90% Pd and 40% Pfa) against low-metal anti-personnel mines at an Army test site seeded with typical clutter. In some cases, university-developed pattern classification and neural network algorithms have reduced the Pfa by a factor of two against these clutter sets.

Tripwire-operated fragmentation mines form a major hazard to deminers, especially where the wire is concealed in vegetation which the deminer has to clear. A similar detection problem is encountered with wire-operated IEDs. This paper discusses some possible technologies for tripwire detection, including acoustic, thermal imaging, and electromagnetic methods, and proposes a diversity approach for use in a reliable, small, inexpensive and long-range detector which can find tripwires completely obscured by vegetation. The results of some experimental measurements using this approach are presented. These suggest that the diversity approach gives a better performance than a pulseinduction detector used alone.

The Humanitarian Demining Research and Development Program of the US Army RDECOM CERDEC Night Vision and Electronic Sensors Directorate (NVESD), under the direction of the Office of Assistant Secretary of Defense for Special Operations and Low-Intensity Conflict (OASD SO/LIC) and with participation from the International Test and Evaluation Program (ITEP) for humanitarian demining, conducted an in-country field evaluation of the Handheld Standoff Mine Detection System (HSTAMIDS) in the southern African country of Namibia. Participants included the US Humanitarian Demining Team of NVESD; ITEP personnel from several member countries; deminers from two non-governmental organizations in Angola, Menschen Gegen Minen (MgM) and HALO Trust; and CyTerra Corporation. The primary objectives were to demonstrate the performance of the U.S. Army's newest handheld multisensor mine detector, the HSTAMIDS, to the performance of the metal detector being used by local demining organizations and also to assess the performance of deminers using the HSTAMIDS after limited experience and training.

We are developing a new landmine detection sensor (ALIS) which is equipped with a metal detector and a GPR. Although this is a hand-held system, we can record the metal detector and GPR signal with the sensor position information. Therefore, signal processing for 2-D signal image is possible. For the metal detector, we apply cross-correlation algorism for sharpening the image and estimation of the depth of the target. For GPR signal, we can apply migration algorithm, which drastically reduce the clutter and we can obtain 3-D image of the buried targets. At first, linear interpolation and cubic interpolation are used respectively to deal with the problem of random data position. Comparing results, we find the image quality of two kinds of interpolations is almost same. Then the migration is used to refocus the scattered signals and improve the image quality for reconstructed landmine image. ALIS demonstration were held in Afghanistan in December 2004 and other countries including Egypt and Croatia in 2005. After some demonstrations and evaluation, we received many useful suggestions. Using these advises, we have modified the ALIS and it is now more easy to use. In this paper, we describe the latest characteristics of the ALIS and summarize its operation.

We developed a hand-held landmine detection sensor system, ALIS (Advanced Landmine Imaging System), combined with a metal detector and GPR (Ground penetrating radar). The system has a CCD camera attached on the sensor handle and can record the MD and GPR signal with the sensor position information. Therefore, it can offer the visual MD image and GPR image, which is used to define targets. But because ALIS is a hand-held system, the sensor position is random when it is operated in the field by human being. Also GPR normally suffers from very strong clutter. To deal with these problems, the interpolation is a common choice for both MD and GPR to create grid data set firstly and migration was used to improve the quality of GPR image. But generally the interpolation can not improve the quality of data set, although it can offer grid data set for visualization. Also for 3D GPR data set, it will consume much processing time. In fact, the migration can not only improve the quality of GPR data but also interpolate data to offer grid data set. It is a kind of 2.5D interpolation and just uses related data in the diffraction trajectory surface. So it can offer directly the visual GPR image and save the processing time. We will discuss two procedures for GPR, interpolation + migration or only migration, in this paper. Lastly, we also will report some results of evaluation test in 2006 February in Croatia.

This paper describes the trials of the MINEHOUND^TM dual sensor, land mine detector carried out in Cambodia, Bosnia and Angola. MINEHOUND^TM has been developed for use in humanitarian demining as a means of improving the efficiency of clearance operations. The trials were sponsored by the UK Department for International Development (DFID). ERA Technology Ltd conducted the trials, which were monitored by staff drawn from the countries participating in the International Test and Evaluation Programme (ITEP) for humanitarian de-mining. Experienced deminers from the Mines Advisory Group (MAG) and Norwegian Peoples Aid (NPA) used the pre-production units in live minefields. The objectives of the trial were: 1. To record information on the performance of MINEHOUND^TM when used in a live minefield. 2. To determine the reduction in False Alarm Rate (FAR) that could be achieved using a dual sensor mine detector. The trials were conducted in three mine-affected countries for a period of eight weeks per country; the programme of trials ran from July 2005 to December 2005, with an additional smaller trial in late February 2006. The results of the trials showed that MINEHOUND^TM achieved 100% detection of the mines encountered and an improvement in FAR of better than 5:1 compared with a basic metal detector. The trials enabled optimisation of the production design and clearly demonstrated that new technology can be brought to humanitarian clearance operations in a safe and controlled manner. As a result of the highly successful trials, Vallon and ERA will produce the MINEHOUND^TM (Type number VMR1) starting in Q3 of 2006.

Spectral features generated from GPR measurements have proven to be effective for the discrimination between landmine and clutter objects. Spectral features are extracted from the energy density spectrum estimated from the GPR data at an alarm location. The quality of the energy density spectrum is highly affected by the ground reflection. For deeply buried plastic landmines, the residual ground response can degrade the energy density spectrum significantly. This paper proposes the use of an estimated depth of the target to improve the estimation of its energy density spectrum, for the purpose of increasing the performance in the detection of deeply buried landmines as well as the discrimination between landmines and clutter objects.

This paper presents and compares two established methods for the automatic detection of landmines in ground penetrating radar (GPR) data. B-scan data of standard GPR targets and simulant landmines were collected from indoor sand and soil lanes. The images were pre-processed and the least squares method and the Hough transform were applied to objects of interest for the detection of hyperbolic signatures in the data. One drawback of the Hough transform is that it can be computationally expensive as it requires a search in 4-D space for hyperbolic shapes. In this case, it has been simplified so only a search in 1-D space is required, however this simplification did result in some missed detections.

In this work we explore and compare several statistical pattern recognition techniques for classification and identification of buried landmines using both electromagnetic induction and ground penetrating radar data. In particular we explore application of different feature extraction approaches to the problem of landmine/clutter classification in blind- and known- ground truth scenarios using data from the NIITEK ground penetrating radar and the Vallon EMI sensor as well as the CyTerra GPR and Minelab EMI sensors. We also compare and contrast the generalization capabilities of different kernels including radial basis function, linear, and direct kernels within the relevance vector machine framework. Results are presented for blind-test scenarios that illustrate robust classification for features that can be extracted with low computational complexity.

We consider in this paper the improvement of side-attack mine detection by performing confidence level fusion with data collected from vehicle-mounted forward-looking IR and GPR (FL-GPR) sensors. The mine detection system is vehicle based, and has both IR and FL-GPR sensors mounted on the top of the vehicle. The IR images and FL-GPR data are captured as the vehicle moves forward. The detections from IR images are obtained from the Scale-Invariant Feature Transform (SIFT) and Morphological Shared-Weight Neural Networks (MSNN) depending on target characteristics, and those from FL-GPR are derived from the FL-GPR SAR images through object-tracking. Since the IR and FL-GPR alarms do not occur at the same location, the fusion process begins with each IR alarm and looks at the nearby FL-GPR alarms with confidences weighted by values that are inversely proportional to their distances to the IR alarm. The FL-GPR alarm with the highest weighted confidence is selected and combined with the IR confidence through geometric mean. An experimental dataset collected from a government test site is used for performance evaluation. At the highest Pd and comparing with IR only, fusing IR and FL-GPR yields a reduction of FAR by 26%. When the Hough transform is applied to reject the IR alarms that have irregular shapes, the fusion results provides a reduction of FAR by 35% at the highest Pd.

A proliferation of the number and variety of sensors for the landmine detection problem has created the need for a sensor manager that is able to intelligently task and coordinate the operation of a suite of landmine sensors. Previous work has developed a framework for sensor management that takes into account the context of the landmine detection problem. The sensor manager searches for N targets in a grid using M multimodal sensors by seeking to maximize the expected information gain. The probabilities of detection and false alarm of the sensors are assumed to be known and are used in the sensor manager calculations. However, in a real-world landmine detection setting, the performance characteristics of the sensors will in fact be unknown. Uneven and irregular ground, vegetation, unanticipated clutter objects, even bad weather - all these can affect the performance of a landmine sensor. This paper examines the effects of uncertainty in the probabilities of detection and false alarm on the performance of the previously presented sensor manager and further examines the performance effects of properly and improperly modeling this uncertainty. Performance is, naturally, found to be adversely affected by uncertainty. However, it is demonstrated that properly modeling the uncertainty present in the problem helps to recover some of the performance that is lost through the introduction of uncertainty.

Landmine sensor technology research has proposed many types of sensors. Some of this technology has matured and can be implemented in sensor arrays that scan for landmines. Other technologies show great promise for distinguishing landmines from clutter, but are more practical to implement on a point-by-point basis as confirmation sensors. This work looks at the problem of scheduling confirmation sensors. Three sensors are considered for their ability to distinguish between landmines and clutter. A novel sensor scheduling algorithm is employed that learns an optimal policy for applying confirmation sensors based on reinforcement learning. A performance gain is realized in both probability of correct classification and processing time. The processing time savings come from not having to deploy all sensors for every situation.

Handheld sensors are commonly used to assist in landmine location and removal. A number of computer systems aimed at assisting humans in discriminating between buried mines and other objects have been developed. Each such system requires some protocol that involves sweeping the sensor over a region of ground using some set of patterns to search for objects (detection) and determine the nature of those objects (discrimination). The work reported here is an effort to determine an acceptable sweep pattern for mine/nonmine discrimination that provides good performance while still being simple for the operator to use. The paper describes a series of data collections and case studies employing a combined radar and metal detection sensor. The system was evaluated first using a robotic operator and later human operators. We discuss the application of a supervised learning system discriminator to each data set, and evaluate discrimination performance. We found that using a relatively simple sweep pattern, computer algorithms can achieve better discrimination performance than an expert human operator, and that (at least up to ten sweeps) our computer algorithm performs better with more sweeps over target.

Previous large-scale blind tests of anti-tank landmine detection utilizing the NIITEK ground penetrating radar indicated the potential for very high anti-tank landmine detection probabilities at very low false alarm rates for algorithms based on adaptive background cancellation schemes. Recent data collections under more heterogeneous multi-layered road-scenarios seem to indicate that although adaptive solutions to background cancellation are effective, the adaptive solutions to background cancellation under different road conditions can differ significantly, and misapplication of these adaptive solutions can reduce landmine detection performance in terms of PD/FAR. In this work we present a framework for the constrained optimization of background-estimation filters that specifically seeks to optimize PD/FAR performance as measured by the area under the ROC curve between two FARs. We also consider the application of genetic algorithms to the problem of filter optimization for landmine detection. Results indicate robust results for both static and adaptive background cancellation schemes, and possible real-world advantages and disadvantages of static and adaptive approaches are discussed.

In the past, many researchers have approached the "Hyperspectral-imagery-anomaly-detection" problem from the point of view of classical detection theory. This perspective has resulted in the development of algorithms like RX (Reed-Xiaoli) and the application of processing techniques like PCA (Principal Component Analysis) and ICA (Independent Component Analysis--algorithms and techniques that are based primarily on statistical and probabilistic considerations. In this paper we describe a new anomaly detection paradigm based on an adaptive filtering strategy known as "signal subspace processing". The signal-subspace-processing (SSP) techniques on which our algorithm is based have yielded solutions to a wide range of problems in the past (e.g. sensor calibration, target detection, and change detection). These earlier applications, however, utilized SSP to relate reference and test signals that were collected at different times. For our current application, we formulate an approach that relates signals from one spatial region in a hyperspectral image to those from a nearby spatial region in the same image. The motivation and development of the technique are described in detail throughout the course of the paper. We begin by developing the signal subspace processing anomaly detector (SSPAD) and proceed to illustrate how it arises naturally from the adaptive filtering formulation. We then compare the algorithm with existing anomaly-detection schemes, noting similarities and differences. Finally, we apply both the SSPAD and various existing anomaly detectors to a hyperspectral data set and compare the results via receiver operating characteristic (ROC) curves.

In this paper, we propose an expectation maximization (EM) trained interacting multiple model (IMM) abrupt change detector for land mine detection applications. The proposed EM algorithm learns the parameters of the different models in real time without requiring a priori information on either the number of models or the model parameters. Using the real ground penetrating radar (GPR) data, the learning performance of the EM-IMM technique is analyzed and commented upon. Numerical receiver operating characteristics (ROC) analysis and detected images indicate that the proposed EM-IMM based abrupt change detector has a better detection and imaging performance than the conventional Kalman filter for land mine detection applications.

Scale Space algorithms, most notably the Scale Invariant Feature Transform (SIFT) are used to produce robust local features from a single training image or a small number of samples. These features can then be computed from, and matched to, test imagery. An object detection/recognition algorithm is built around this matching process. The aim of this project is to locate a particular class of side-attack landmines, PARMs, in SWIR video using a SIFT-based detector. The detection of PARMs is accomplished by generating SIFT keypoints for the training images and each frame of the video and then determining a model of keypoint matches that represent the scale, orientation, and location of PARMs in the video with a high degree of certainty. Once a SIFT match between the training image set and a frame of video is found, a new hypothesis about the location, relative orientation, and scale of a PARM in the video sequence is created. Each new keypoint match is then assigned a confidence-based score. If this new keypoint match is compatible with the geometrical model of any previously generated hypothesis, the keypoint match's score, scaled by a fuzzy membership function, is added to the score of that hypothesis. An alarm is triggered once the score of a hypothesis reaches a predefined threshold. Results of the side-attack landmine detection are presented.

We describe an efficient approach for finding probable target areas quickly with a minimal number of Ground Penetrating Radar (GPR) measurements. Since a potential GPR target creates a hyperbolic signature in the space-time domain, our approach uses the time delay differences from consecutive GPR A-Scan data to estimate the location of the apex of the hyperbolic signature, thus locating a target. This apex prediction method uses many fewer measurements than a full backprojection algorithm. Regions of low target probability are determined using a Neyman-Pearson detection approach in order to eliminate redundant measurements. In this regard, our approach is especially suitable as a pre-screener: other sensors that are more accurate, but require more measurement time, can then be applied only to high probability-of-target areas to corroborate results, differentiate between targets, or provide more accurate location measurements. Compared to a standard backprojection algorithm more signal-to-noise ratio (SNR) is needed to achieve similar detection performance. This SNR loss can be reduced by using a more conservative algorithm which reduces the step size of the GPR antenna. Results from experimental data collected at a model mine field at the Georgia Institute of Technology show that target positions can be found accurately using less than 10% of the measurements utilized by conventional imaging algorithms.

The AN/PSS-14 (a.k.a. HSTAMIDS) has been tested for its performance in South East Asia (Thailand), South Africa (Namibia) and in November of 2005 in South West Asia (Afghanistan). The system has been proven effective in manual demining particularly in discriminating indigenous, metallic artifacts in the minefields. The Humanitarian Demining Research and Development (HD R&D) Program has sought to further improve the system to address specific needs in several areas. One particular area of these improvement efforts is the development of a mine detection/discrimination improvement software algorithm called Region Processing (RP). RP is an innovative technique in processing and is designed to work on a set of data acquired in a unique sweep pattern over a region-of-interest (ROI). The RP team is a joint effort consisting of three universities (University of Florida, University of Missouri, and Duke University), but is currently being led by the University of Florida. This paper describes the state-of-the-art Region Processing algorithm, its implementation into the current HSTAMIDS system, and its most recent test results.

This paper describes an algorithm for land mine detection in GPR data that uses edge histogram descriptors for feature extraction and fuzzy K-Nearest Neighbors (K-NN) for confidence assignment. First, an LMS algorithm for anomaly detection is used to focus attention and identify candidate signatures that resemble mines. Second, translation invariant features are extracted based on spatial distribution of edges in the 3-D GPR signatures. Specifically, each 3-D signature is divided into sub-signatures, and the local edge distribution for each sub-signature is represented by a histogram. To generate the histogram, local edges are categorized into five types: vertical, horizontal, diagonal, anti-diagonal, and non-edges. Next, the training signatures are clustered to identify prototypes. The main idea is to identify few prototypes that can capture the variations of the signatures within each class. These variations could be due to different mine types, different soil conditions, different weather conditions, etc. Fuzzy memberships are assigned to these representatives to capture their degree of sharing among the mines and false alarm classes. Finally, fuzzy K-NN based rules are used to assign a confidence value to distinguish true detections from false alarms. The proposed algorithm is applied to data acquired from three outdoor test sites at different geographic locations.

In order to extract accurate quantitative information out of Ground Penetrating Radar (GPR) measurement data, one needs to solve a nonlinear inverse problem. In this paper we reformulate this into a nonlinear least squares problem which is non convex. Solving a non-convex optimization problem requires a good initial estimation of the optimal solution. In this paper we use a three step method to solve the just described non-convex problem. In a first step the qualitative solution of the linearized problem is found to obtain the detection and support of the subsurface scatterers. For this first step Synthetic Aperture Radar (SAR) and MUltiple SIgnal Classification (MUSIC) are proposed and compared. The second step consists out of a qualitative solution of the linearized problem to obtain a first guess for the material parameter values of the detected objects. The method proposed for this is Algebraic Reconstruction Theorem (ART), which is an iterative method, starting from the initial value, given by the first step, and improving on this until an optimum is achieved. The final step then consists out of the solution of the nonlinear inverse problem using a variational method. The paper starts with a discussion of the GPR inverse problem and continues with a short description of the used methods (SAR, MUSIC, ART and adjoint method). Finally an example is given based on simulated data and some conclusions are drawn.

Understanding sorption mechanisms of Explosive Related Chemicals (ERCs) in subsurface environments is essential in predicting their fate and transport, since sorption onto the soil may reduce the ERC concentration in the liquid and gas phase, thus affecting its subsequent detection. This project is studying the equilibrium and non-equilibrium sorption of 2,4-Dinitrotoluene (DNT) onto soils under different conditions such as temperature and soil type (sandy soil, clayey soil). The sorption behavior of DNT in tropical soils samples from Isabela, P.R. under saturated conditions is currently being studied at 25±2°C. Kinetic sorption experiments showed that equilibrium was achieved after approximately 30 hours for sand and approximately 12 hours for clay. Equilibrium studies in the range of 2-14 mg/L of DNT followed a Freundlich isotherm. These results demonstrate it is not appropriate to assume a linear relationship between the amount of solute sorbed onto the solid and the concentration of the solute as may not always be the case for concentrations outside a certain range that will be dependent on the type of soil and organic matter content. This might result in under or over estimation of adsorption, and thus the fate and transport and eventual detection of DNT. At low DNT concentrations (less than 6 mg/L for sand and less than 8 mg/L for clay), the isotherms were quasilinear. In this range, a linear adsorption constant, K_d value of 0.0172 L/Kg and 1.46 L/Kg for sand and clay, respectively, showed there was more affinity of the DNT for clay than for sand.

Deep Ultra Violet Raman Spectroscopy (DUV-RS) is an emerging tool for vibrational spectroscopy analysis and can be used in Point Detection mode to detect explosive components of landmines and Improvised Explosive Devices (IED). Interactions of explosives with different substrates can be measured by using quantitative vibrational signal shift information of scattered Raman light associated with these interactions. In this research, grounds were laid for detection of explosives using UV-Raman Spectroscopy equipped with 244 nm laser excitation line from a 488 nm frequency doubled Coherent FreD laser. In other experiments, samples of 2,4-DNT were allowed to interact with Ottawa Sand and were studied using DUV-RS. Characteristic vibrational signals of energetic compounds were analyzed in the ranges: 400-1200 cm^-1, 1200-1800 cm^-1, and 2800-3500 cm^-1. In addition these Raman spectra were compared with dispersive spectra that were acquired using Raman Microscopy equipped with 514.5 nm (VIS) 785 nm (NIR) and 1064 nm (NIR) excitation lasers.

This paper presents the development and testing of a three-dimensional laboratory-scale soil tank system for modeling ERC fate and transport under controlled, but variable environmental conditions in partially saturated soil. The system incorporates a rainfall simulator, variable light (visible and UV), temperature and relative humidity components, and a 3D SoilBed capable of simulating several boundary and initial conditions. Experimental work indicate that water and solute transport is highly influenced by interrelated environmental and boundary conditions. The presence of light and higher system temperatures induces greater water drainage and solute fluxes. During infiltration, hydraulic heads increase at faster rates under no light exposure suggesting greater water and solute retention. The spatial and temporal distribution of hydraulic heads during rainfall events is not uniform and flow patterns reflect preferential paths. Transport of conservative solutes closely follows water flow patterns, and reflects the influence of variable and interrelated environmental factors on spatial and temporal concentration distribution. These experiments show that interrelated environmental factors must be taken into account to accurately predict the distribution of chemicals near the soil-atmosphere surface. They demonstrate that non-reactive solutes are highly influenced by variation in hydraulic, advective, and dispersive processes induced by changes in environmental conditions. Greater impacts are expected for reactive and semi-volatile solutes such as ERCs. In such case, fate and transport will also be affected by variations in soptive, gas transport, and degradation processes.

Landmines and other buried explosive devices pose in an immense threat in many places of the world, requiring large efforts on detection and neutralization of these objects. Many of the available detection techniques require the presence of chemicals near the soil-atmospheric surface. The presence of explosive related chemicals (ERCs) near this surface and their relation to the location of landmines, however, depends on the source characteristics and on fate and transport processes that affect their movement in soils. Fate and transport processes of ERC is soils may be interrelated with each other and are influenced by chemical characteristics and interrelated soil and environmental factors. Accurate detection of ERCs near the soil surface must, therefore, take into the variability of ERC concentration distributions near the soil surface as affected by fate and transport processes controlled interrelated environmental factors. To effectively predict the concentration distributions of ERCs in soils and near soil surfaces, it is necessary to have good understanding of parameters values that control these processes. To address this need, field lysimeters have been designed and developed at the University of Puerto Rico, Mayaguez .This paper presents the design of two field lysimeter used to study the fate and transport behavior of ERC in the field subjected to varying uncontrolled subtropical environmental conditions in two different soils. Both lysimeters incorporate pressure and concentration sampling ports, thermocouples, and a drainage system. Hydrus-2D was used to simulate soil moisture and drainage in the lysimeter for average environmental conditions in the study for the two soils used. The field lysimeters allow collection and monitoring of spatial and temporal ERC concentrations under variable, uncontrolled environmental conditions.

2,4,6-Trinitrotoluene (TNT) has a number of specific properties that make it a nearly ideal explosive for military applications. It is relatively stable with respect to non desired detonation, easy to store and handle and has a high explosive power. A broad variety of landmines contain TNT as the main explosive charge. There are several methods currently used to detect buried landmines, both physically and chemically. The goal of this work is develop new methods for detecting TNT in contact with soil, based on Chemical Point Detection methodologies. FT-IR spectroscopy is used to provide information about identity and composition of compounds in very small samples or small heterogeneities in large samples. The main objective of this work is to study the vibrational behavior of TNT when in contact with soil that contains argillaceous minerals, specifically of the group of the smectites. Literature indicates that clays of this group present certain characteristics leading to affinity to nitroaromatic compounds, such as TNT. The clay used in this investigation was saturated with potassium cations to increase the adsorption of TNT on clay. The study includes the exposure of Clay/TNT mixtures to a series of environmental variables, which include: variation of alkalinity and acid content of the mixtures, variation of temperature, addition of water and explosive mass fraction in the mixture. Visible changes of color in the K-clay-TNT or Na-clay-TNT mixtures were observed but without displaying vibrational changes in highly basic clays.

Detection of Explosive Related Chemicals (ERCs) emanating from landmines is strongly influenced by fate and transport processes in variably saturated soils, which are affected by many environmental factors. To study the fate and transport behavior of ERCs in soils, it is necessary to conduct experiments is physical models designed for this purpose. Sampling and analysis of the chemicals in soil water and air is one of the most important components of this design. In this project, air and water sampling devices and methods for sampling TNT and DNT in a tropical sandy soil were studied. Different stainless steel porous samplers were evaluated to determine sampling volumes and efficiencies. Results show that they can be used with proper extraction vacuum and pore size depending on soil water content. The stainless steel porous samplers show no or little effect of the sampling efficiencies of TNT and DNT.

Infrared Spectroscopy is a well established tool for standoff detection of chemical agents in military applications. Vibrational IR spectroscopic analysis can also be used in Chemical Point Detection mode and to the arena of explosives identification and detection when energetic compounds are in contact with soil. PETN is an important nitroaliphatic explosive for military applications. Due to its intrinsic explosive power, it can be used in laminar form or mixed with RDX to manufacture Semtex plastic explosive and in the fabrication of Improvised Explosive Devices (IEDs). This investigation focused on the study of spectroscopic signatures of PETN in contact with soil. For this study, clay was mixed in different proportions with PETN. Detection of the vibrational signatures of PETN constitutes the central part of the investigation. The mixtures were submitted to the effect of water, acid and alkaline solutions, heat and deep UV light (234 nm) in order to establish the effect on these environmental parameters on the vibrational signatures of the explosive in the mixtures. The results reveal that the characteristic bands of PETN are highly persisted, degraded only by extreme conditions of UV radiation and exposure to high temperature for prolonged time. These results could be used in the development of sensitive sensors for detection of landmines, and improvised explosives devices (IDEs).

This paper presents the development of a simulation tool to facilitate the exploration and evaluation of design tradeoffs for an Unmanned Aerial Vehicle (UAV) based minefield detection system. Mine and minefield performance estimates and design tradeoffs are obtained using explicit evaluation of detection statistics simulated under different sensors, minefield layout scenarios, and mission specific constraints. The simulated mine and minefield level performance results are compared with analytical results where available. Design tradeoffs are studied in terms of different sensor and mission profile parameters such as signal to clutter ratio, target size, field-of-regard, and detection algorithms. The analytical relationship and simulated results of mine and minefield detection performance based on these parameters are presented. Different metrics for evaluating minefield performance and their influences on design tradeoffs are discussed, and suggestions are made.