Forward-looking Ground Penetrating Radar Research Articles

Forward-looking ground penetrating radar image reconstruction using fully connected neural network

One way to identify targets buried in the ground is to use forward-looking ground penetration radars. The two components of speed and accuracy are very important in reconstructing the images obtained from this method. Distinguishing the target from the Earth clutter is one of the main challenges of this type of imaging. This article proposes a novel fast algorithm that reconstructs scattered images created by forward-looking ground-penetrating radar. The linear back-projection algorithm is used for fast image reconstruction. Because our problem is nonlinear ill-posed inverse, errors and distortions are compensated by a deep fully connected neural network. The difference between received and reconstructed signal is the input of the neural network, and output is the error between reconstructed and actual coefficients of the Interested region. Our assumption is the region of interest has a few electromagnetic scatterers. Because of types of soils and non-stationary situations, it is difficult to estimate the probability density function of noise plus clutter. Therefore, we modified order statistic constant false alarm rate to detect shallowly buried scatterers. The simulation results show that the algorithm improves signal-to-noise and clutter ratio in output and detects objects with a constant probability of false alarm in a fast way.

Robust Copula-Based Detection of Shallow-Buried Landmines With Forward-Looking Radar

We propose a technique for landmine detection using forward-looking ground-penetrating radar. The detector is applied to radar images obtained from multiple viewpoints of the region of interest and is based on a robust version of the likelihood ratio test (LRT). We incorporate the statistical dependence between multiview images into the test via copula-based model. The test is designed to maximize the worst-case performance over all feasible pairs of target and clutter distributions, thereby eliminating the need for strong assumptions about the image statistics. We evaluate the detection performance of the proposed technique for different copula functions representing the dependence structure. Using electromagnetic modeled data of shallow-buried targets under varying ground surface roughness profiles, we demonstrate the superiority of the robust copula-based detector over existing parametric and robust LRT approaches designed under the assumption of statistical independence of multiview images.

Analysis and Validation of a Hybrid Forward-Looking Down-Looking Ground Penetrating Radar Architecture

Ground Penetrating Radar (GPR) has proved to be a successful technique for the detection of landmines and Improvised Explosive Devices (IEDs) buried in the ground. In the last years, novel architectures for safe and fast detection, such as those based on GPR systems onboard Unmanned Aerial Vehicles (UAVs), have been proposed. Furthermore, improvements in GPR hardware and signal processing techniques have resulted in a more efficient detection. This contribution presents an experimental validation of a hybrid Forward-Looking–Down-Looking GPR architecture. The main goal of this architecture is to combine advantages of both GPR architectures: reduction of clutter coming from the ground surface in the case of Forward-Looking GPR (FLGPR), and greater dynamic range in the case of Down-Looking GPR (DLGPR). Compact radar modules working in the lower SHF frequency band have been used for the validation of the hybrid architecture, which involved realistic targets.

Forward-Looking Ground-Penetrating Radar: Subsurface target imaging and detection: A review

Detection of shallow-buried, in-road threats using a forward-looking (FL) ground-penetrating radar (GPR) system has attracted significant research interest in the last decade. An FL-GPR mounted on a moving platform can provide standoff target detection and imaging. This enables real-time sensing and situation awareness over large ground areas. The main challenge facing this sensing technology is high false-alarm rates due to scattering arising from air–ground interface roughness and subsurface clutter.

Minimax Robust Landmine Detection Using Forward-Looking Ground-Penetrating Radar

We propose a robust likelihood-ratio test (LRT) to detect landmines and unexploded ordnance using a forward-looking ground-penetrating radar. Instead of modeling the distributions of the target and clutter returns with parametric families, we construct a band of feasible probability densities under each hypothesis. The LRT is then devised based on the least favorable densities within the bands. This detector is designed to maximize the worst-case performance over all feasible density pairs and, hence, does not require strong assumptions about the clutter and noise distributions. The proposed technique is evaluated using electromagnetic field simulation data of shallow-buried targets. We show that, compared to detectors based on parametric models, robust detectors can lead to significantly reduced false alarm rates, particularly in cases where there is a mismatch between the assumed model and the true distributions.

Coherence-Factor-Based Rough Surface Clutter Suppression for Forward-Looking GPR Imaging

We present an enhanced imaging procedure for suppression of the rough surface clutter arising in forward-looking ground-penetrating radar (FL-GPR) applications. The procedure is based on a matched filtering formulation of microwave tomographic imaging, and employs coherence factor (CF) for clutter suppression. After tomographic reconstruction, the CF is first applied to generate a “coherence map” of the region in front of the FL-GPR system illuminated by the transmitting antennas. A pixel-by-pixel multiplication of the tomographic image with the coherence map is then performed to generate the clutter-suppressed image. The effectiveness of the CF approach is demonstrated both qualitatively and quantitatively using electromagnetic modeled data of metallic and plastic shallow-buried targets.

Multiview Synthetic Aperture Ground-Penetrating Radar Detection in Rough Terrain Environment: A Real-Time 3-D Forward Model

A major problem with forward-looking ground-penetrating radar (FLGPR) detection of buried explosive threats is the scattering from the rough ground surface. The authors have previously developed a real-time 3-D algorithm for single-view emulation of synthetic aperture FLGPR scattering from rough terrain at low grazing angles. This article extends the method to the real-time 3-D simulation of a multiview moving platform transmitter/receiver array moving as fast as 15 km/h. The need to perform the new calculation at every frame for the moving antenna array platform is shown to be unnecessary, with the requirement relaxed to once for only a single frame at the center of 10–20 m forward interrogation range, leading to about 50 times faster computation. The computation of the scattered waves comprising surface clutter is reduced for all moving frames to a mere multiplication of three matrices: a pre-computed impulse response matrix of rough terrain, a pre-computed correction matrix of moving frames, and a matrix characterizing the transmitting input signal. The method is evaluated via 3-D Monte Carlo simulation for various rough surface parameters, and its applicability for subsurface scattering reconstruction to characterize the buried threat objects is shown. For a vehicle-mounted FLGPR detection system, this leads to a significant saving of computation resources. Our developed algorithm provides 3-D modeling of rough terrain scattering for lossy and frequency-dispersive soil and compares well with full-wave finite-difference frequency-domain (FDFD) method computation.

Bistatic Landmine and IED Detection Combining Vehicle and Drone Mounted GPR Sensors

This work proposes a novel Ground Penetrating Radar (GPR) system to detect landmines and Improvised Explosive Devices (IEDs). The system, which was numerically evaluated, is composed of a transmitter placed on a vehicle and looking forward and a receiver mounted on a drone and looking downwards. This combination offers both a good penetration and a high resolution, enabling the detection of non-metallic targets and mitigating the clutter at the air–soil interface. First, a fast ray tracing simulator was developed to find proper configurations of the system. Then, these configurations were validated using a full wave simulator, considering a flat and a rough surface. All simulations were post-processed using a fast and accurate Synthetic Aperture Radar (SAR) algorithm that takes into account the constitutive parameters of the soil. The SAR images for all configurations were compared, concluding that the proposed contribution greatly improves the target detection and the surface clutter reduction over conventional forward-looking GPR systems.

Real-Time Modeling of Forward-Looking Synthetic Aperture Ground Penetrating Radar Scattering From Rough Terrain

Ground penetrating radar (GPR) is a viable tool for fast and high fidelity detection of concealed explosive threats. The radar effectiveness is limited by scattering from rough terrain which considerably obscures the buried target response. To calculate the rough ground scattering, a 3-D full-wave algorithm such as finite-difference frequency-domain (FDFD) method is required but is often prohibitive for multiple frames when the GPR antennas are distant from the target region. This paper presents a real-time 3-D modeling of a moving platform forward-looking GPR scattering from rough terrain located at great electrical distances from the GPR antenna. For a synthetic aperture, the computational domain of the focal region is reduced to a very small subset of the entire observed volume, and the surface clutter is computed via a mere multiplication of a precomputed impulse response matrix of the rough ground with the matrix characterizing the GPR transmitting signal. For a vehicle-mounted GPR detection system, this results in a significant reduction of complexity and saving of computation resources. The effectiveness of the algorithm is evaluated through an implementation of 3-D Monte Carlo simulation for various rough surface parameters. Our developed model compares well with the direct FDFD results, and can be used for lossy and frequency-dispersive soils.

Fast Image Reconstruction in Forward-Looking GPR Using Dual $\ell _1$ Regularization

In the forward-looking ground-penetrating radar (GPR) imaging scenario, detection and classification performance is limited due to inherent constraints placed on the sensor configuration as well as limitations associated with the conventional image formation process. Image reconstruction using sparsity inducing regularization has been demonstrated as a promising technique to overcome these limitations but at a significant computational cost. In this paper, we present a dual $\ell _1$ regularization technique, which is consistent with the expectation that objects are sparsely populated but extended with respect to the pixel grid. Furthermore, we propose a series of preprocessing steps, which produce a greatly reduced data size and computational cost. Demonstrations of the proposed approach are provided using experimental data and results are given in terms of signal to background ratio, image resolution, and relative computation time.

High-Resolution Multiple-Input Multiple-Ouput FLGPR Imaging

Multiple-input multiple-output forward-looking ground-penetrating radar (GPR) systems can be used to detect landmines. To enhance its performance, two data-dependent imaging algorithms, based on the sparse learning via iterative minimization (SLIM) and sparse covariance-based estimation (SPICE) techniques for high-resolution imaging, applied to the time-domain GPR data, were previously developed. Time-domain SLIM (TD-SLIM) and time-domain SPICE (TD-SPICE) yield higher resolution and lower sidelobe compared with data-independent approaches such as delay-and-sum and recursive sidelobe minimization. However, the two data-dependent imaging algorithms are computationally expensive. To provide both improved landmine detection performance and significantly reduced computational time compared with the TD-SLIM and TD-SPICE algorithms, we propose herein the frequency-domain SLIM algorithm.

A Large Comparison of Feature-Based Approaches for Buried Target Classification in Forward-Looking Ground-Penetrating Radar

Forward-looking ground-penetrating radar (FLGPR) has recently been investigated as a remote sensing modality for buried target detection (e.g., landmines). In this context, raw FLGPR data is beamformed into images and then computerized algorithms are applied to automatically detect subsurface buried targets. Most existing algorithms are supervised, meaning they are trained to discriminate between labeled target and non-target imagery, usually based on features extracted from the imagery. A large number of features have been proposed for this purpose, however thus far it is unclear which are the most effective. The first goal of this work is to provide a comprehensive comparison of detection performance using existing features on a large collection of FLGPR data. Fusion of the decisions resulting from processing each feature is also considered. The second goal of this work is to investigate two modern feature learning approaches from the object recognition literature: the bag-of-visual-words and the Fisher vector for FLGPR processing. The results indicate that the new feature learning approaches outperform existing methods. Results also show that fusion between existing features and new features yields little additional performance improvements.

Adaptive Detection of Low-Signature Targets in Forward-Looking GPR Imagery

We present an image-domain adaptive likelihood ratio tests (LRT) detector for low-signature target detection in forward-looking ground-penetrating radar. We exploit multiview tomographic images of the scene under investigation to iteratively adapt the statistics of the targets and clutter arising from the interface roughness. Using numerical electromagnetic data, it is shown that the proposed adaptive LRT detector provides significantly lower false-alarm rates compared with its nonadaptive counterpart while providing comparable detection performance.

Multiview Imaging for Low-Signature Target Detection in Rough-Surface Clutter Environment

Forward-looking ground-penetrating radar (FL-GPR) permits standoff sensing of shallow in-road threats. A major challenge facing this radar technology is the high rate of false alarms stemming from the vulnerability of the target responses to interference scattering arising from interface roughness and subsurface clutter. In this paper, we present a multiview approach for target detection in FL-GPR. Various images corresponding to the different views are generated using a tomographic algorithm, which considers the near-field nature of the sensing problem. Furthermore, for reducing clutter and maintaining high cross-range resolution over the imaged area, each image is computed in a segmentwise fashion using coherent integration over a suitable set of measurements from multiple platform positions. We employ two fusion approaches based on likelihood ratio tests detector to combine the multiview images for enhanced target detection. The superior performance of the multiview approach over single-view imaging is demonstrated using electromagnetic modeling data.

Forward-Looking Radar Imaging: A Comparison of Two Data Processing Strategies

This paper aims at comparing the conventional backpropagation algorithm and a microwave tomographic approach in the framework of forward-looking ground-penetrating radar imaging. The reconstruction capabilities of both methods are investigated in terms of achievable resolution limits for a sensing configuration similar to the Synchronous Impulse Reconstruction radar developed at the U.S. Army Research Laboratory. Furthermore, reconstruction results obtained by processing full-wave simulated data are shown with the goal of comparing the capability of both methods to detect buried and surface targets in scenarios emulating realistic operation conditions.

Forward-Looking Ground-Penetrating Radar via a Linear Inverse Scattering Approach

An inverse scattering approach has been designed to process data gathered by means of forward-looking ground-penetrating radar systems. Such an inverse approach exploits a linear model of the scattering phenomenon and is able to account for the half-space 2-D geometry. In this frame, a theoretical study on the achievable reconstruction capabilities is provided with the aim to estimate the range and cross-range resolution limits and gain indications about the issue of how to design the measurement configuration. Finally, numerical and experimental examples are provided to assess the effectiveness of the approach.

Performance analysis of forward-looking GPR ultra-wideband antennas for buried object detection

We are currently developing a Stepped-Frequency Radar (SFR) which utilizes a custom-made uniform linear array of 16 Vivaldi notch receive antennas and two Transverse Electromagnetic (TEM) horn transmit antennas. The SFR has an operating band of 300–2000 MHz, and a minimum frequency step-size of 1 MHz. The custom-made TEM horn antennas are used for the transmission of the SFR’s ultra-wideband (UWB) spectrum. This paper discusses a comparison analysis between a commercially available UWB antenna and the currently used TEM horns. Gain, Voltage Standing Wave Ratio (VSWR), and antenna pattern measurements for each antenna are presented. The antennas were also tested for their ability to detect buried targets in a simple stepped-frequency radar system using a network analyser as a transmitter and receiver. An analysis of the gain, VSWR, beamwidth, and measured data from radar test of each antenna was performed, providing insights into each antenna’s performance on the SFR’s ability to detect buried targets. The information provided in this paper will be useful to the radar community in exploring developmental standoff detection solutions for military applications such as obscured target detection of obstacles and explosive hazards.

High-Resolution Imaging for Impulse-Based Forward-Looking Ground Penetrating Radar

Forward-Looking Ground Penetrating Radar (FLGPR) has multiple applications, one of which includes its use for detecting landmines and other buried improvised explosive devices (IEDs). The standard method for generating synthetic aperture radar (SAR) images for this radar is the backprojection (BP) algorithm, which has poor resolution and high sidelobe problems. In this paper, we consider using the Sparse Iterative Covariance-based Estimation (SPICE) algorithm and the Spare Learning via Iterative Minimization (SLIM) algorithm for generating sparse high-resolution images for FLGPR. A pre-processing step, which involves an orthogonal projection of the received data onto a subspace related to the region of interest is performed, for decreasing the dimension of the data and for clutter reduction. The SLIM and SPICE algorithms are user-parameter free, and are capable of providing SAR images with improved resolution. We also use the well-known CLEAN approach for imaging based on a proposed signal model in the time domain. We show using simulated data that the SPICE and SLIM algorithms provide higher resolution than CLEAN and the standard BP. Imaging using real data collected via the Synchronous Impulse Reconstruction (SIRE) radar, a multiple-input multiple-output (MIMO) FLGPR radar developed by the Army Research Laboratory (ARL), is also presented and used for analysis.

Ultra‐wide band antenna designs and numerical system modelling for forward‐looking GPR

ABSTRACTThis paper presents ultra‐wide band (UWB) horn feeder and parabolic reflector antenna designs, which are considered highly suitable for forward‐looking vehicle‐mounted ground‐penetrating radar (GPR) systems and time‐domain numerical modelling of impulse GPR signal radiation, wave propagation and target scattering. In this context, the partial dielectric‐loaded Vivaldi shaped TEM horn (PDVA), TEM horn fed double‐ridged horn (PDTEM‐RHA) and their compact versions are investigated, designed simulated and measured. An offset parabolic reflector antenna prototype with a compact PDTEM‐RHA feeder is proposed to reach hyper‐wide band impulse radiation performances from 500 MHz up to 15 GHz for a multi‐band forward‐looking GPR operation that can provide deep subsurface detection with high‐resolution imaging, due to a big advantage of narrowed beam widths. The antenna gain and radiation pattern performances are demonstrated with measurement results.

Sparse MIMO Array Forward-Looking GPR Imaging Based on Compressed Sensing in Clutter Environment

This paper presents a sparse multiple-input and multiple-output (MIMO) array and sparse frequency ground-penetrating radar (GPR) imaging scheme based on compressed sensing (CS). Since the targets of interest for GPR are usually sparse, the number of the MIMO array elements and frequencies can be reduced using CS theory. Thus, the system complexity and data acquisition time can be reduced accordingly. Considering the serious clutter in forward-looking GPR, we propose two methods for the CS reconstruction in clutter environment. The first one is a clutter suppression preprocessing method, which can effectively suppress the azimuth clutter and short range clutter outside the reconstruction region and significantly improve the reconstruction result. The second one is to determine the regularization parameter for the CS reconstruction in clutter environment. We refer to this reconstruction process as basis pursuit declutter. The proposed imaging scheme can produce pointlike and less cluttered images of sparse targets using fewer array elements and frequencies. Results from simulated data, trihedral reflector, and real buried land mine experimental data are presented to show the validity of the proposed methods. The experimental data are acquired by the vehicle-mounted stepped-frequency forward-looking ground-penetrating virtual aperture radar, which is designed and developed by the National University of Defense Technology.