Search

This paper aims at detecting and characterizing inclusions in concrete structures by inverting ground-penetrating radar (GPR) data. First, the signal is preprocessed using the principal component analysis (PCA) and then used to train an artificial neural network (ANN). The GPR data consists of 1200 time steps. Using PCA, the data can be compressed to 286 dimensions without losing any information. Moreover, with 99.99% of the original variance the data needs only 139 dimensions. This dimensional reduction makes the ANN training easier and faster. The ANN were trained to find the buried inclusions characteristics--depth and radii--considering a nonhomogenous host medium by inverting the preprocessed data. The results show that the expected maximum error was kept under 1%, which is a remarkable result, since the host medium is nonhomogenous. Index Terms--Artificial neural network (ANN), buried objects, ground-penetrating radar (GPR), inverse problem.

In this paper, the development of an ultra-wideband (UWB) array-based time-domain radar sensor for near-field imaging is described. The radar sensor is designed to be used within a vehicle-mounted multisensor system for humanitarian demining. The main novelty of the radar lies in the system design with a single transmitter and multichannel receiver. Design of the UWB antenna array is also novel. The radar produces 3-D images of subsurface by 1-D mechanical scanning. The imaging capability of the radar is realized via electronic steering of the receive antenna footprint in a cross-scan direction and synthetic aperture processing in an along-scan direction. Imaging via footprint steering allows for a drastic increase in the scanning speed. Index Terms--Ground penetrating radar (GPR), landmine detection, near-field imaging, ultra-wideband (UWB) antenna array.

There has been significant amount of study on the use of ground-penetrating radar (GPR) for land-mine detection. This paper presents our analysis of GPR data collected at a U.S. Army test site using a new approach based on frequency subband processing. In this approach, from the radar data that have over 2.5 GHz of bandwidth, we compute separate radar images using the one wide (2 GHz) and four narrow (0.6 GHz) frequency subbands. The results indicate that signals for different frequency subbands are significantly different and give very different performance in land-mine detection. In addition, we also examine a number of features extracted from the GPR data, including magnitude and local-contrast features, ratio between copolarization and cross-polarization signals, and features obtained using polarimetric decomposition. Feature selection procedures are employed to find subsets of features that improve detection performance when combined. Results of land-mine detection, including performance on blind test lanes, are presented. Index Terms--Feature extraction, feature selection, frequency subbands, ground-penetrating radar (GPR), land-mine detection.

A common approach to training neural network classifiers in a supervised learning setting is to minimize the mean-square error (mse) between the network output for each labeled training sample and some desired output. In the context of landmine detection and discrimination, although the performance of an algorithm is correlated with the mse, it is ultimately evaluated by using receiver operating characteristic (ROC) curves. In general, the larger the area under the ROC curve (AUC), the better. We present a new method for maximizing the AUC. Desirable properties of the proposed algorithm are derived and discussed that differentiate it from previously proposed algorithms. A hypothesis test is used to compare the proposed algorithm to an existing algorithm. The false alarm rate achieved by the proposed algorithm is found to be less than that of the existing algorithm with 95% confidence. Index Terms--Area under the ROC curve (AUC), ground-penetrating radar (GPR), landmine detection, pattern recognition.

Full-resolution 3-D ground-penetrating radar (GPR) imaging of the near surface should be simple and efficient. Geoscientists, archeologists, and engineers need a tool capable of generating interpretable subsurface views at centimeter-to-meter resolution of field sites ranging from smooth parking lots to rugged terrain. The authors have integrated novel rotary laser positioning technology with GPR into such a 3-D imaging system. The laser positioning enables acquisition of centimeter accurate x, y, and z coordinates from multiple small detectors attached to moving GPR antennas. Positions streaming with 20 updates/s from each detector are fused in real time with the GPR data. The authors developed software for automated data acquisition and real time 3-D GPR data quality control on slices at selected depths. Industry standard (SEGY) format data cubes and animations are generated within an hour after the last trace has been acquired. Such instant 3-D GPR can be used as an on-site imaging tool supporting field work, hypothesis testing, and optimized excavation and sample collection in the exploration of the static and dynamic nature of the shallow subsurface. Index Terms--Data acquisition, data processing, ground-penetrating radar (GPR), imaging, lasers, position measurement, terrain mapping.

This paper proposes a new inverse method for microwave-based subsurface sensing of lossy dielectric objects embedded in a dispersive lossy ground with an unknown rough surface. An iterative inversion algorithm is employed to reconstruct the geometry and dielectric properties of the half-space ground as well as that of the buried object. B-splines are used to model the shape of the object as well as the height of the rough surface. In both cases, the control points for the spline function represent the unknowns to be recovered. A single-pole rational transfer function is used to capture the dispersive nature of the background. Here, the coefficients in the numerator and denominator are the unknowns. The approach presented in this paper is based on the state-of-the-art semianalytic mode matching forward model, which is a fast and efficient algorithm to determine the scattered electromagnetic fields. Numerical experiments involving two-dimensional geometries and TM incident plane waves demonstrate the accuracy and reliability of this inverse method. Index Terms--B-splines, dispersive media, ground-penetrating radar (GPR), inversion methods, nonlinear optimization, rough surface, subsurface sensing.

The factorization method (FM) has been applied to measurement data from a multistatic ground-penetrating radar operating in close proximity to the ground, which was used in a measurement campaign on the Joint Research Centre mine test lane in Ispra, Italy. This paper is targeted toward a future hand-held demining system. The according space limits restrict an independent positioning of transmit and receive antennas. Hence, very small multistatic datasets are obtained, representing a difficult case for the reconstruction with the FM. Index Terms--Angular diversity, antipersonnel mine, discrete complex image method (DCIM), factorization method (FM), Green's function, ground-penetrating radar (GPR), linear sampling method, multistatic radar, stratified media.

A ground-penetrating radar (GPR) antenna system is modeled using the three-dimensional finite-difference time-domain technique. This paper investigates some basic geometrical shapes for planar dipoles to find what shape gives the best performance for GPR applications. The antenna is resistor loaded and shielded by a rectangular conducting cavity to suit the application. The effect of adding a wave-absorbing coat to the shield is also studied. Furthermore, a genetic algorithm is used to optimize the cavity height and the resistor values. Index Terms--Absorbing media, buried-object detection, dipole antennas, genetic algorithms.

A new parametric inversion technique to locate cylindrical structures has been developed for borehole radar cross-hole measurements. The technique calculates prediction errors, focusing on curve shapes of first-arrival times, and explicitly uses known parameters on a target. Two schemes to compare the measured first-arrival time curves with the calculated ones are proposed. One is by taking the errors between curve gradients of measured and calculated first-arrival times, and the other is by taking the cross correlation between calculated first-arrival times and measured data. The schemes do not need to select the true first-arrival times, which makes this inversion technique robust and easy to use. This technique is validated with synthetic data sets modeling a metallic pipe in homogeneous and heterogeneous medium models. By both models, the inversion technique is able to retrieve the exact location of the pipe, though it employs a rather simple forward model, which can be solved by only geometrical calculations. The technique is applied to measured data sets, and it successfully estimates a pipe location in good agreement with known information. The inversion is also used for an air-filled subsurface cavity with a very simplified forward model, which considers only Snell's law at the cavity-subsurface media interface. It is also able to locate the cavity at a similar location estimated by other conventional techniques. Index Terms--Borehole radar, cross-hole, ground-penetrating radar (GPR), inverse problem, pipe detection, tunnel detection.

Thin surface layers of high-permittivity material (e.g., water-saturated soil) can cause a pronounced dispersion of ground-penetrating radar waves. The dispersion characteristics depend on the permittivity and thickness of the effective surface waveguide and the permittivity of the material below it. As for the analogous seismic case, the values of these parameters can be estimated by applying a scheme that includes calculating the phase-velocity spectra, picking dispersion curves from the spectra, and inverting the dispersion curves using a combined local and global minimization procedure. Here, this scheme is extended by incorporating higher order modes. Explicit expressions for the cutoff frequencies of the fundamental and higher order transverse electric (TE) and transverse magnetic (TM) modes are derived. These expressions demonstrate that the cutoff frequency decreases as the thickness and/or permittivity of the waveguide increase and/or the permittivity of the lower half-space decreases. In addition, the TE cutoff frequencies are shown to be generally lower than the TM cutoff frequencies. Numerical modeling of realistic models demonstrates the presence of the higher order modes with the correct cutoff frequencies. Application of the modified inversion scheme to synthetic and field data demonstrates its efficacy in providing the required physical property information. In particular, better-constrained models are obtained by including higher order modes in the inversion. Index Terms--Dispersion, ground-penetrating radar (GPR), inversion, transverse electric (TE), transverse magnetic (TM), waveguide.

We show how a single reflector antenna with a multimode feedhorn can be used in a ground moving target indication (GMTI) radar. In particular, we demonstrate the simultaneous detection and estimation of angular location of a ground moving target via adaptive cancellation of ground clutter.

A two-stage algorithm for landmine detection with a ground penetrating radar (GPR) system is described. First, 3-D data sets are processed using a computationally inexpensive pre-screening algorithm which flags potential locations of interest. These flagged locations are then passed to a feature-based processer which further discriminates target-like anomalies from naturally occurring clutter. Current field trial (over 6500 square meters) and blind test results (over 39000 square meters) are presented and these show at least an order of magnitude improvement over other radar system-based detection algorithms on the same test lanes.

Most of the current forward-looking ground-penetrating radar (FLGPR) systems use conventional delay-and-sum (DAS) based methods to form radar images for detection of the target (such as a landmine). However, DAS is a data-independent approach which is known to suffer from low resolution and poor interference and clutter rejection capability. We present a data-adaptive imaging approach for FLGPR image formation based on APES (amplitude and phase estimation) and rank-deficient RCB (robust Capon beamforming). Due to the data-adaptive nature of both APES and RCB, our approach has better resolution and much better interference and clutter rejection capability than the standard DAS-based imaging methods. The excellent performance of the proposed method is demonstrated using experimental data collected via two FLGPR systems recently developed by PSI (Planning Systems, Inc.) and SRI (Stanford Research Institute).

We consider landmine detection using forward-looking ground penetrating radar (FLGPR). The two main challenging tasks include extracting intricate structures of target signals and adapting a classifier to the surrounding environment through learning. Through the time-frequency (TF) analysis, we find that the most discriminant information is TF localized. This observation motivates us to use the over-complete wavelet packet transform (WPT) to sparsely represent signals with the discriminant information encoded into several bases. Then the sequential floating forward selection method is used to extract these components and thereby a neural network (NNW) classifier is designed. To further improve the classification performance and deal with the problem of detecting mines in an unconstraint environment, the AdaBoost algorithm is used. We integrate the feature selection process into the original AdaBoost algorithm. In each iteration, AdaBoost identifies the hard-to-learn examples and a new set of features which provide the specific discriminant information for these hard samples is extracted adaptively and a new classifier is trained. Experimental results based on measured data are presented, showing that a significant improvement on the classification performance can be achieved.

Neural networks have been applied to landmine detection from data generated by different kinds of sensors. Real-valued neural networks have been used for detecting landmines from scattering parameters measured by ground penetrating radar (GPR) after disregarding phase information. This paper presents results using complex-valued neural networks, capable of phase-sensitive detection followed by classification. A two-layer hybrid neural network structure incorporating both supervised and unsupervised learning is proposed to detect and then classify the types of landmines. Tests are also reported on a benchmark data. Index Terms--Hybrid artificial neural networks, landmine classification, landmine detection, scattering parameters.

In this paper, the basic design of an adaptive ground penetrating radar antenna is introduced. The antenna is able to adapt its input impedance to a variation in the antenna elevation and soil type to keep reflections at the antenna's terminal minimum. As a result, energy transfer from the generator to the antenna is maximized, which in turn maximizes the energy radiated by the antenna into the ground for different antenna elevations and soil types. The antenna is based on a wire bow-tie structure with variable flare angle for adjusting the antenna's input impedance. The flare angle variation is realized by short-circuiting the gaps separating the wires from the feed point of the antenna, for which electronic switching devices such as PIN diodes could be used to allow fast and convenient control of the antenna's flare angle. Index Terms--Adaptive antenna, bow-tie antenna, ground penetrating radar (GPR) antenna, wire antenna.

Ground-penetrating radar (GPR) has been proposed as an alternative to classical electromagnetic induction techniques for the landmine detection problem. The Wichmann/Niitek system provides a good platform for novel GPR-based antitank mine detection and classification algorithm development due to its extremely high SNR. When the GPR sensor is mounted on a moving vehicle, the target signatures are hyperbolas in a time-domain data record. The goal of this work is to extract useful features that exploit this knowledge in order to improve target detection. The algorithms can be divided into two steps: feature extraction and classification. Preprocessing is also considered to remove both stationary effects and nonstationary drift of the data and to improve the contrast of the desired hyperbolas. The algorithm is evaluated using real data over primarily plastic antitank mines collected with a fielded GPR sensor at a government test site. Index Terms--Antitank landmine detection, ground-penetrating radar (GPR).

The accuracy of a migration image of ground-penetrating radar (GPR) depends strongly on the accuracy of permittivity distribution determined from multioffset data. This paper proposes a migration velocity analysis method using a genetic algorithm (GA). The objective function is defined as the summation of normalized zero-delay cross correlation of all common-image point gathers. Under the assumptions that the media are blockwise and that the permittivity of each block can be expressed as a polynomial with limited terms, all coefficients of the permittivity function of each block, which maximize the objective function, are determined by migration velocity analysis method with GA. Prestack migration is performed by a reverse-time migration method based on Maxwell's equations solved by the finite-difference time-domain method with a perfectly matched layer absorbing boundary conditions. The migration velocity analysis method is applied to synthetic common-transmitter datasets to test the method. Then, the velocity analysis and prestack migration method are applied to field data. From the distribution of dielectric constant obtained from the field data, water content is derived, and the depth of a water aquifer is deduced from the water content distribution and a migration stack profile. Index Terms--Genetic algorithm (GA), ground-penetrating radar (GPR), migration velocity analysis, prestack migration, water content.

This paper presents a newly developed high-order discontinuous Galerkin time-domain (DGTD) method for solving Maxwell's equations in linear dispersive media with UPML boundary treatment. A unified formulation is derived for linear dispersive media of Debye type and the artificial material in the UPML regions with the help of auxiliary differential equations. The DGTD employs finite-element-type meshes, and uses piecewise high-order polynomials for spatial discretization and Runge-Kutta method for time integrations. Arbitrary high-order accuracy can be obtained for scattering of various objects in dispersive media. After validating the numerical convergence of the DGTD method together with the second-order Yee's scheme, we apply this new method to the ground-penetrating radar for the detection of buried objects in a lossy half space. Index Terms--Discontinuous Galerkin time-domain (DGTD), dispersive media, ground-penetrating radar (GPR), perfectly matched layer (PML).

The use of ground-penetrating radar (GPR) techniques at high frequencies for historical building investigations requires an accurate knowledge of the material complex permittivity. A new measurement system is presented, which by exploiting a resonance phenomenon, yields accurate values of permittivity and resistivity also in the case of low-loss materials. These parameters are obtained from an amplitude-only measurement of the transmission response curve of a rectangular waveguide cavity, loaded by the sample under test. The setup requires the availability of brick-shaped samples, which can be obtained easily in the case of rock materials. Permittivities and resistivities were measured in the frequency range 1.1-1.7 GHz on samples of marbles, sandstones, and limestones. The measurement method and the experimental results ate presented and discussed. Index Terms--Complex permittivity measurement, dielectric losses, ground-penetrating radar (GPR), rock materials.

A ground-penetrating radar (GPR) with a suspended horn antenna was used to obtain continual measurement of near-surface soil water content dynamics within a well-defined footprint. Water contents inferred from radar surface reflectivity (SR) were in agreement with gravimetric measurements from the top 1-cm soil layer. This sensitivity to top 1 cm was confirmed by comparison of SR with deeper measurements (1-5 cm) obtained by time-domain reflectometry. Measurements over sand and silt loam showed the role of soil type in controling near-surface dynamics of soil water due to evaporation and drainage processes. GPR with a horn antenna provided insights into physical processes that could affect larger scale radar platforms (air- or spaceborne) and enabled verification of radar measurements at well-defined spatial scales and detailed temporal resolutions not available by other radar remote sensing systems. Index Terms--Dielectric measurements, ground-penetrating radar (GPR), remote sensing, soil water content.

In this paper, we present a new technique whereby phase variation signatures are used to profile two-dimensional (2-D) rough surfaces and to discern shallowly buried, small, low-contrast landmines under a flat ground. The method has been tested using data measured over a composite surface containing two rough dielectric surface patches, and over a flat ground under which small, low-contrast antipersonnel (AP) landmines are shallowly buried. The results show that the phase-based technique is capable of profiling rough surfaces and of detecting small, low-contrast landmines with different internal structures buried underneath a flat ground. Index Terms--Ground penetrating radar, phase measurement, phase variation signatures, rough surfaces.

An adaptive framework is presented for ultra-wideband ground penetrating radar imaging of shallowly-buried low-contrast dielectric objects in the presence of a moderately rough air-soil interface. The proposed approach works with sparse data and relies on recently developed Gabor-based narrow-waisted quasi-ray Gaussian beam algorithms as fast forward scattering predictive models. First, a nonlinear inverse scattering problem is solved to estimate the unknown coarse-scale roughness profile. This sets the stage for adaptive compensation of clutter-induced distortion in the underground imaging problem, which is linearized via Born approximation and subsequently solved via variouspixel-based and object-based techniques. Numerical simulations are presented to assess accuracy, robustness and computational efficiency for various calibrated ranges of problem parameters. The proposed approach has potential applications to antipersonnel land mine remediation. Index Terms--Curve evolution, Gaussian beams, ground penetrating radars, rough surfaces, short pulses, total variation.

Optimized versions of frequency-wavenumber (F-K) migration methods are introduced to better focus ground-penetrating radar {{;PR} data in applications of shallow subsurface object localization, e.g., landmine remediation. Migration methods are based on the wave equation and operate by backpropagating the received data into the earth so as to localize buried objects. Traditional F-K migration is based on an underlying assumption that the wavefields propagate in a homogeneous medium. The presence of a rough air-ground interface in the GPR case degrades the localization ability. To overcome this problem in the context of the F-K algorithm, we introduce lateral variations in the velocity of waves in the medium. An optimization approach is employed to choose that velocity function that results in a well-focused image where an entropy-like criterion is used to quantify the notion of focus. Extension of the basic method to lossy medium is also described. The utility of these techniques is demonstrated using field data from a number of GPR systems. Index Terms--Complex-velocity frequency-wavenumber (F-K) migration, dispersive medium, frequency-wavenumber (F-K) migration, minimum entropy optimization, Tikhonov regularization, varimax norm.