Asphalt thickness estimation using ground-penetrating radar (GPR) is an important application in transportation infrastructure inspections. Thin asphalt measurement needs to use super-resolution or full-wave methods to process the GPR signals with overlapped echoes. We propose a model for antenna reflection coefficient in the presence of multilayer medium by omitting considerations of evanescent waves and interactions between the antenna and the medium. We formulated it within the spectral domain, enabling the explicit incorporation of the antenna’s radiation pattern. This model is about 100 times faster than the existing full-wave models. In this model, we calculate the plane wave expansion (PWE) of the antenna radiated field. Ignoring the evanescent waves, the PWE of the radiated field is calculated from the far-field pattern of the antenna. Simulation and measurement results confirm the validity and effectiveness of this model in asphalt thickness measurement.

Ground-penetrating radar (GPR) has become indispensable for subsurface reconstruction, especially in complex, multilayered infrastructure environments. Efficient and accurate estimation of permittivity and layer thickness, crucial for understanding these scenes, presents significant challenges. This article explores the integration and interaction between the back projection (BP) algorithm and full-wave inversion (FWI) to optimize reconstruction outcomes. We propose a ray-based refraction method and a high-order moment (HOM)-oriented local BP estimation, each aimed at improving efficiency from the critical perspectives of “delay” and “summation” in the BP algorithm, respectively, with a focus on the target layer. Crucially, the BP estimation results provide a reliable initial model for FWI, reducing the risk of local minima and decreasing the iteration count for more precise parameter estimation. FWI also compensates for the BP algorithm’s limitations in estimating parameters in nontarget layers. Furthermore, full-wave modeling (FWM) mitigates antenna effects prior to parameter estimation, thereby enhancing accuracy. During reconstruction, the BP algorithm primarily generates target imaging, while FWI provides detailed information about layer interfaces. This synergistic approach leverages the complementary strengths of both algorithms: the BP algorithm captures spatial information depicted by the targets observed in B-scans, while FWI incorporates detailed electromagnetic wave propagation along layer interfaces presented in A-scans via radar equations. Experiments systematically analyze the accuracy and effectiveness of the proposed approach using ideal simulation models, sandbox laboratory data, and road data from the Belgian Road Research Centre (BRRC) facilities. This comprehensive evaluation underscores the approach’s substantial potential in advanced geophysical surveys and environmental research.

Radar sensing technologies offer several advantages over other gesture input modalities, such as the ability to reliably sense human movements, a reasonable deployment cost, insensitivity to ambient conditions such as light, temperature, and the ability to preserve anonymity. These advantages come at the price of high processing complexity mainly due to the spatio-temporal variations of gesture articulation performed by different people. Deep learning methods, such as CNN-LSTM and 3D CNN-LSTM, have a high potential to recognize radar-based gestures but usually require hundreds or thousands of labeled training samples and high processing power. Asking a lot of people to acquire a lot of gestures is particularly tedious and tiring to the point of being unrealistic. To overcome these challenges, we propose FORTE, a hand gesture recognition with few samples based on an optimized CNN architecture working on pre-processed raw data. Using a k=5-fold cross-validation, we define and compare three alternative CNNs for recognizing hand gestures acquired in a semi-mobile context of use with a portable radar attached to a smartphone. The best CNN reaches an accuracy of 94.96% with a precision of 95.92% and a recall of 96.03% for a dataset composed of solely 5 participants producing 2 samples for 20 classes covering 1 pointing, 2 pantomimic, 3 iconic, and 14 semaphoric gestures. We suggest some implications for designing radar-based gestures and we discuss the limitations of this approach.

Microwave radars bring many benefits to mid-air gesture sensing due to their large field of view and independence from environmental conditions, such as ambient light and occlusion. However, radar signals are highly dimensional and usually require complex deep learning approaches. To understand this landscape, we report results from a systematic literature review of ( N =118) scientific papers on radar sensing, unveiling a large variety of radar technology of different operating frequencies and bandwidths and antenna configurations but also various gesture recognition techniques. Although highly accurate, these techniques require a large amount of training data that depend on the type of radar. Therefore, the training results cannot be easily transferred to other radars. To address this aspect, we introduce a new gesture recognition pipeline that implements advanced full-wave electromagnetic modeling and inversion to retrieve physical characteristics of gestures that are radar independent, i.e., independent of the source, antennas, and radar-hand interactions. Inversion of radar signals further reduces the size of the dataset by several orders of magnitude, while preserving the essential information. This approach is compatible with conventional gesture recognizers, such as those based on template matching, which only need a few training examples to deliver high recognition accuracy rates. To evaluate our gesture recognition pipeline, we conducted user-dependent and user-independent evaluations on a dataset of 16 gesture types collected with the Walabot, a low-cost off-the-shelf array radar. We contrast these results with those obtained for the same gesture types collected with an ultra-wideband radar made of a vector network analyzer with a single horn antenna and with a computer vision sensor, respectively. Based on our findings, we suggest some design implications to support future development in radar-based gesture recognition.

Microwave radar sensors in human-computer interactions have several advantages compared to wearable and image-based sensors, such as privacy preservation, high reliability regardless of the ambient and lighting conditions, and larger field of view. However, the raw signals produced by such radars are high-dimension and relatively complex to interpret. Advanced data processing, including machine learning techniques, is therefore necessary for gesture recognition. While these approaches can reach high gesture recognition accuracy, using artificial neural networks requires a significant amount of gesture templates for training and calibration is radar-specific. To address these challenges, we present a novel data processing pipeline for hand gesture recognition that combines advanced full-wave electromagnetic modelling and inversion with machine learning. In particular, the physical model accounts for the radar source, radar antennas, radar-target interactions and target itself, i.e.,, the hand in our case. To make this processing feasible, the hand is emulated by an equivalent infinite planar reflector, for which analytical Green’s functions exist. The apparent dielectric permittivity, which depends on the hand size, electric properties, and orientation, determines the wave reflection amplitude based on the distance from the hand to the radar. Through full-wave inversion of the radar data, the physical distance as well as this apparent permittivity are retrieved, thereby reducing by several orders of magnitude the dimension of the radar dataset, while keeping the essential information. Finally, the estimated distance and apparent permittivity as a function of gesture time are used to train the machine learning algorithm for gesture recognition. This physically-based dimension reduction enables the use of simple gesture recognition algorithms, such as template-matching recognizers, that can be trained in real time and provide competitive accuracy with only a few samples. We evaluate significant stages of our pipeline on a dataset of 16 gesture classes, with 5 templates per class, recorded with the Walabot, a lightweight, off-the-shelf array radar. We also compare these results with an ultra wideband radar made of a single horn antenna and lightweight vector network analyzer, and a Leap Motion Controller.

Drone-borne ground-penetrating radar (GPR) has proven to be efficient and promising for high-resolution topsoil moisture mapping at the field scale. However, in practice, the radar incident angle may change as a function of the terrain slope and unstable flying conditions. In this respect, we analyzed the effect of radar incident angle on full-wave inversion for soil permittivity characterization. Specifically, using the frequency-domain radar equation, antenna characteristic functions with different incident angles were determined for a horn antenna as an example. Then, numerical analyses and field measurements were performed to quantify the errors on the retrieved permittivity resulting from incident angle drifts. In agreement with the radar equation concept, we observed that the angle- and frequency-dependent global transmission functions of the antenna correspond well to its radiation pattern. Numerical and field results show that errors in the permittivity estimation can be very significant as a function of the antenna radiation pattern and the incident angle. Nevertheless, if the incident angle is known and the characteristic antenna functions are determined as a function of the incident angle, full-wave modeling and inversion remain accurate, provided that the signal-to-noise ratio remains sufficient.

In this study, we analysed low-frequency drone-borne ground-penetrating radar (GPR) and full-wave inversion for soil electrical conductivity mapping. Indeed, in the lowest GPR frequency ranges, the soil surface reflexion coefficient depends more on the soil electrical conductivity than on its permittivity. Numerical experiments were conducted within the frequency range 15-45 MHz to analyse parameter sensitivities, the well-posedness of the inverse problem as well as the depth of sensitivity. The results show that the soil surface reflexion is significantly more sensitive to the soil electrical conductivity than the soil permittivity. Therefore, the conductivity can be retrieved using full-wave inversion within this frequency range, with a characterization depth varying from 0.5 to 1 m, depending on the soil properties. Yet, the permittivity also affects the results and should be accounted for in the inversion strategy. Field measurements were performed using low-frequency drone-borne radar with a 5-metre half-wave dipole antenna, and EMI measurements with different depth sensitivities were conducted for comparison. Kriging interpolation was used to get maps from measurement points. The soil conductivity maps obtained by the proposed GPR and EMI are compliant in terms of absolute values and spatial patterns. This study demonstrated the capacity of low-frequency drone-borne GPR for fast, field-scale soil electrical conductivity mapping.

Ground-penetrating radars operating at a stand-off distance from the probed domain are attractive diagnostic tools in several contexts. Despite notable advantages offered in terms of noninvasiveness and flexibility, the usage of stand-off configurations makes the imaging process more difficult. Indeed, an accurate reconstruction of the subsurface/hidden scene requires accounting for the presence of the air-soil interface into the scattering model. This article deals with a comparison of three different linear models for describing the scattering phenomenon under a stand-off configuration, where the presence of the layered medium is explicitly accounted for in the scattering model. The first model is rigorous and based on the spectral domain representation of the incident field and Green's function for the half-space scenario. The second model exploits a ray representation of the wave propagation from the radar to the target and vice versa by accounting for the refraction phenomenon at the interface. The third model is an approximated one, where the layered nature of the scenario is taken into account by means of an equivalent wavenumber, which allows a straightforward evaluation of the incident field and Green's function. The resolution performance of each inversion approach is assessed by means of a numerical analysis. Moreover, data referred to an experiment carried out in controlled conditions are processed and reconstructions are provided to support the comparative analysis among the three imaging approaches.

Antenna effects alter the detection of buried objects during ground-penetrating radar (GPR) surveys. In this paper, we propose a novel approach based on full-wave inversion to filter out antenna effects from GPR data. The approach, which is exact for locally planar layered media, resorts to a recently developed electromagnetic model that takes advantage of an intrinsic, closed-form solution of Maxwell's equations to describe the antenna-medium system. As any multilayered medium can be reduced to a half-space medium with effective, frequency-dependent, global reflection coefficients, the method consists in inverting the radar data to retrieve a frequency-dependent, half-space complex conductivity. Converted into the time domain, this quantity represents the filtered radar image. We validated the approach through numerical simulations and laboratory experiments with pipes buried in a sandbox. The results demonstrated the validity of the concept and showed that the filtered radar images include only medium reflections, which means an easier interpretation in terms of medium structures. Antenna radiation pattern effects are, however, not removed. This physically based approach including the full-wave antenna model appears to be very promising for improved subsurface imaging and provides the basis for multifrequency GPR data fusion as the source is inherently normalized.

Horizontally stratified media are commonly used to represent naturally occurring and man-made structures, such as soils, roads, and pavements, when probed by ground-penetrating radar (GPR). Electromagnetic (EM) wave scattering from such multilayered media is dependent on the roughness of the interfaces. In this paper, we developed a closed-form asymptotic EM model considering random rough layers based on the scalar Kirchhoff-tangent plane approximation (SKA) model that we combined with planar multilayered media Green's functions. In order to validate our extended SKA model, we conducted simulations using a numerical EM solver based on the finite-difference time-domain (FDTD) method. We modeled a medium with three layers-a base layer of perfect electric conductor (PEC) overlaid by two layers of different materials with rough interfaces. The reflections at the first and at the second interface were both well reproduced by the SKA model for each roughness condition. For the reflection at the PEC surface, the extended SKA model slightly overestimated the reflection, and this overestimation increased with the roughness amplitude. Good agreement was also obtained between the FDTD simulation input values and the inverted root mean square (rms) height estimates of the top interface, while the inverted rms heights of the second interface were slightly overestimated. The accuracy and the performances of our asymptotic forward model demonstrate the promising perspectives for simulating rough multilayered media and, hence, for the full waveform inversion of GPR data to noninvasively characterize soils and materials.

Multifrequency and multioffset ground-penetrating radar data acquisition modes are used to maximize the information content and parameter retrieval capabilities. However, they also increase the computational cost dedicated to the inversion procedure. In this paper, the impact of the number of frequencies and the multistatic configurations on the information retrieval capabilities is investigated through the response surface topographies of the objective functions. We resort to a full-wave-inversion procedure and a recently developed electromagnetic model which takes advantage of a closed-form solution of Maxwell's equations to describe the antenna-medium system. We show with numerical and laboratory experiments the possibility of reducing the number of frequencies from several hundreds to one or several tens of components without affecting the information retrieval capabilities. We also show through several scenarios that the presence of a perfect electrical conductor increases the number of frequencies required to ensure an acceptable retrieval of the subsurface properties whereas the conductivity of the first layer and the relative permittivity of the second layer do not affect it. The results highlight that information content analyses are important in order to study and optimize data acquisition and inversion procedures, and thereby the computation time.

In this paper, we propose an automated detection algorithm for well- and ill-shaped ground-penetrating radar reflection hyperbolas for complex media, calibrated with human recognition principles. The algorithm detects the apex of the hyperbolas by fitting an analytical function of a hyperbola to the profile edge dots detected with a Canny filter. The existence of a hyperbola is determined using a set of carefully chosen criteria calibrated in order to fit the ambiguity zone for the human brain. The inherent misshapedness of field hyperbolas is further considered by defining a buffer zone around the theoretical hyperbola. First, the method was tested in the laboratory over tree roots and PVC pipes and on field images over tree root systems. Both time- and frequency-domain radars were used on-ground. After around 1-3 min of computation time for 10 000 edge dots in a MATLAB environment (single 1.96-GHz processor), the results showed rates of false alarm and nondetection of maximum 20% and 28%, respectively. In comparison with the semiautomated hyperbola detection provided by a commercial software, these rates were lower. Second, we conducted a sensitivity analysis to estimate the validity of the fitting of a hyperbola equation neglecting the object radius. The fitting was close, but the derivation of the relative permittivity from the analytical equation neglecting the radius led to high errors. In conclusion, owing to the low computational time and its good performances, the proposed algorithm is suitable for complex environments.

Litter is acknowledged to affect microwave remote sensing data over forests, and accurate quantification of litter radiative properties is essential for proper retrieval of forest soil water content and forest canopy biophysical characteristics from satellite or airborne sensors. Furthermore, detailed characterization of forest soil organic horizons is of paramount importance for ecological studies and for carbon cycle and global changerelated studies. Ultrawideband ground-penetrating radar (GPR) data acquired above beech forest litter were used to analyze the effects of litter on the backscattering and investigate the ability of GPR to quantitatively characterize the properties of the forest floor. Radar measurements were performed over different litter configurations (i.e., single layers of recently fallen litter (OL) or of fragmented litter (OF), and combined OL-OF layers) and thicknesses. Radar data were processed through full-wave inversion. Over the considered frequency range (0.8-4.0 GHz), attenuation of the radar signal increased as both frequency and litter thickness increase, due to scattering and dielectric losses. These effects were accounted for in the radar model through frequency dependence of litter apparent electrical conductivity. Good agreement (RMSE = 9.5 × 10-3 m) was observed between estimated and measured litter thicknesses. Radar data inversions provided reliable estimates of litter electromagnetic properties, with relative dielectric permittivity values around 1.2 and 3.9 for OL and OF litters, respectively, and corresponding values of 0.032 and 0.080 Sm-1for litter apparent electrical conductivity at 4.0 GHz. These results show great promise for the use of GPR for noninvasive characterization of forest litter.

In this paper, we experimentally analyzed the feasibility of estimating soil hydraulic properties from 1.4 GHz radiometer and 0.8-2.6 GHz ground-penetrating radar (GPR) data. Radiometer and GPR measurements were performed above a sand box, which was subjected to a series of vertical water content profiles in hydrostatic equilibrium with a water table located at different depths. A coherent radiative transfer model was used to simulate brightness temperatures measured with the radiometer. GPR data were modeled using full-wave layered medium Green's functions and an intrinsic antenna representation. These forward models were inverted to optimally match the corresponding passive and active microwave data. This allowed us to reconstruct the water content profiles, and thereby estimate the sand water retention curve described using the van Genuchten model. Uncertainty of the estimated hydraulic parameters was quantified using the Bayesian-based DREAM algorithm. For both radiometer and GPR methods, the results were in close agreement with in situ time-domain reflectometry (TDR) estimates. Compared with radiometer and TDR, much smaller confidence intervals were obtained for GPR, which was attributed to its relatively large bandwidth of operation, including frequencies smaller than 1.4 GHz. These results offer valuable insights into future potential and emerging challenges in the development of joint analyses of passive and active remote sensing data to retrieve effective soil hydraulic properties.

We present a closed-form equation for intrinsic modeling of near-field electromagnetic induction (EMI) antennas for planar layered media characterization. Resorting to a decomposition of the backscattered EM field into elementary distributions over the antenna aperture, the EMI transmitting and receiving antennas are modeled using infinitesimal magnetic dipoles and field points, and characteristic frequency-dependent global reflection and transmission coefficients. Low-frequency propagation of the EM fields in the medium is described using 3-D planar layered media Green's functions. We performed measurements with a loop antenna situated at different heights, ranging from near-field to far-field conditions, above water of known electrical conductivity to determine its intrinsic properties, and a range of salinity conditions was applied to subsequently validate the proposed model. The EMI system was set up using a vector network analyzer equipped with a prototype EMI antenna specifically designed for this application. The model showed good accuracy for reproducing the observed data, and model inversion provided good estimates of the medium electrical conductivity. Yet, insensitivity of the EMI signal to water electrical conductivity was encountered for low salinity due to the presence of a copper sheet as the bottom boundary condition of the experimental setup. Moreover, the efficiency of the antenna decreased rapidly as antenna height above water surface increases, leading to increasing discrepancies between estimated and measured water electrical conductivity values as the antenna moves away from the water surface. Although some technical improvements are still needed, the proposed approach is promising for quantitative estimation of soil electrical conductivity from EMI data.

Full-wave modeling of ground-penetrating radar (GPR) data using Green's functions for wave propagation in planar layered media and antenna characteristic global reflection and transmission functions for describing far-field antenna effects, including antenna-medium interactions, has shown a great potential for nondestructive characterization of soils and materials. The accuracy of the retrieved parameters in the GPR data inversion depends on the accuracy of the GPR external calibration. In this research we studied the stability and the repeatability of two different GPR systems, namely, frequency- and time-domain systems. A combination of a vector network analyzer and 800-5200 MHz horn antenna was used as a frequency-domain GPR (FD-GPR) whereas a GSSI GPR system using a 900 MHz bowtie antenna was used as a time-domain GPR (TD-GPR). Both GPR systems including their antennas were calibrated several times using measurements with the antennas at different heights over a perfect electric conductor (PEC) in the laboratory as well as over a water layer. In addition, measurements were performed over a thin water layer and a relatively thick sandy soil layer as validating medium. The results showed that the FD-GPR is relatively stable while the TD-GPR presents a significant drift which can be accounted for using corrections based on the air direct-coupling waves (free-space measurements). Water- and PEC-based calibrations provided very similar results for the GPR calibration functions. Inversions for the water layer and the sandy soil layer provided reliable results and showed a high degree of the repeatability for both radar systems. The error on the calibration based on inaccurate antenna heights over PEC showed the significant errors on the inversion results for the directive antenna (horn antenna) but less error for the bowtie antenna. This analysis demonstrated the general validity of the proposed far-field radar modeling approach, not only with respect to frequency and time domain radars but as well with respect to the calibrating medium.

A new near-field radar modeling approach for wave propagation in planar layered media is presented. The radar antennas are intrinsically modeled using an equivalent set of infinitesimal electric dipoles and characteristic, frequency-dependent, global reflection, and transmission coefficients. These coefficients determine through a plane wave decomposition wave propagation between the radar reference plane, point sources, and field points. The interactions between the antenna and layered medium are thereby inherently accounted for. The fields are calculated using 3-D Green's functions. We validated the model using an ultrawideband frequency-domain radar with a transmitting and receiving Vivaldi antenna operating in the range 0.8-4 GHz. The antenna characteristic coefficients are obtained from near- and far-field measurements over a copper plane. The proposed model provides unprecedented accuracy for describing near-field radar measurements collected over a water layer, the frequency-dependent electrical properties of which were described using the Debye model. Layer thicknesses could be retrieved through full-wave inversion. The proposed approach demonstrated great promise for nondestructive testing of planar materials and digital soil mapping using ground-penetrating radar.

We present validation results of a new ground-penetrating radar (GPR) near-field model for determining the electrical properties and correlated water content of a sand using both frequency- and time-domain radars. The radar antennas are intrinsically characterized using an equivalent set of infinitesimal source/field points and characteristic functions of antennas, which were determined using measurements with the antenna at different distances from a copper plane. The antenna radiation was modeled using six source and field points, which was found to be a good compromise between high modeling accuracy and computing efficiency. We validated our model by inverting GPR data to predict the water content of a sand layer subject to seven levels of saturation. A soil dielectric mixing model was integrated into the full-wave GPR inverse modeling to directly estimate the water content and to account for the frequency dependence of the electrical properties. Although the quality of the fit slightly decreased as the antenna approached the sand surface, the results showed a close agreement between measured and modeled data, resulting in accurate estimation of the water content. The average errors of all water content estimates were 0.012 cm3/cm3for the frequency domain and 0.016 cm3/cm3for the time-domain GPR. However, the accuracy reduced when the sand became wet. By performing numerical simulations, we found that it is due to the vertical heterogeneity of soil moisture under the effect of the hydrostatic pressure. We also showed that the GPR inversion with the multilayered soil model could account for this heterogeneity and improved the agreement between the modeled and measured GPR data as well as the accuracy of soil moisture estimation. As for the frequency dependence of the electrical properties, in the frequency ranges of both GPR systems, while the dielectric permittivity was approximately constant, the apparent conductivity exponentially increased with increasing frequency. The success of the calibration and validation in laboratory conditions demonstrates a great potential for practical applications of the radar model, notably for the digital soil mapping and nondestructive testing of materials.

In this study, we experimentally analyze the feasibility of estimating the soil hydraulic properties from L-band radiometer and ground-penetrating radar (GPR) data. L-band radiometer and ultrawideband off-ground GPR measurements were performed above a sand box in hydrostatic equilibrium with a water table located at different depths. The results of the inversions showed that the radar and radiometer signals contain sufficient information to estimate the soil water retention curve and its related hydraulic parameters with a relatively good accuracy compared to time-domain reflectometry estimates. However, an accurate estimation of the hydraulic parameters was only obtained by considering the saturated water content parameter as known during the inversion.

The objective of this study was to evaluate two remote-sensing methods for mapping the surface soil moisture of a bare soil, namely L-band radiometry using brightness temperature and ground-penetrating radar (GPR) using surface reflection inversion. Invasive time-domain reflectometry (TDR) measurements were used as a reference. A field experiment was performed in which these three methods were used to map soil moisture after controlled heterogeneous irrigation that ensured a wide range of water content. The heterogeneous irrigation pattern was reasonably well reproduced by both remote-sensing techniques. For GPR, the effect of roughness was excluded by operating at low frequencies (0.2-0.8 GHz) that were not sensitive to the field surface roughness. For the radiometer, the effect of roughness was accounted for using an empirical model that required calibration with the reference TDR measurements. The root mean square (RMS) error between soil moisture measured by GPR and TDR was 0.038 m3m-3while the RMS error between radiometer (horizontal and vertical polarizations)and TDR-derived soil water content was 0.020 m3m-3. These results suggest that both remote-sensing techniques are promising for field-scale mapping of surface soil moisture over bare soils.

The antenna of a zero-offset off-ground ground-penetrating radar can be accurately modeled using a linear system of frequency-dependent complex scalar transfer functions under the assumption that the electric field measured by the antenna locally tends to a plane wave. First, we analyze to which extent this hypothesis holds as a function of the antenna height above a multilayered medium. Second, we compare different methods to estimate the antenna phase center, namely, 1) extrapolation of peak-to-peak reflection values in the time domain and 2) frequency-domain full-waveform inversion assuming both frequency-independent and -dependent phase centers. For that purpose, we performed radar measurements at different heights above a perfect electrical conductor. Two different horn antennas operating, respectively, in the frequency ranges 0.2-2.0 and 0.8-2.6 GHz were used and compared. In the limits of the antenna geometry, we observed that antenna modeling results were not significantly affected by the position of the phase center. This implies that the transfer function model inherently accounts for the phase-center positions. The results also showed that the antenna transfer function model is valid only when the antenna is not too close to the reflector, namely, the threshold above which it holds corresponds to the antenna size. The effect of the frequency dependence of the phase-center position was further tested for a two-layered sandy soil subject to different water contents. The results showed that the proposed antenna model avoids the need for phase-center determination for proximal soil characterization.

Accurate estimates of surface soil moisture are essential in many research fields, including agriculture, hydrology, and meteorology. The objective of this study was to evaluate two remote-sensing methods for mapping the soil moisture of a bare soil, namely, L-band radiometry using brightness temperature and ground-penetrating radar (GPR) using surface reflection inversion. Invasive time-domain reflectometry (TDR) measurements were used as a reference. A field experiment was performed in which these three methods were used to map soil moisture after controlled heterogeneous irrigation that ensured a wide range of water content. The heterogeneous irrigation pattern was reasonably well reproduced by both remote-sensing techniques. However, significant differences in the absolute moisture values retrieved were observed. This discrepancy was attributed to different sensing depths and areas and different sensitivities to soil surface roughness. For GPR, the effect of roughness was excluded by operating at low frequencies (0.2-0.8 GHz) that were not sensitive to the field surface roughness. The root mean square (rms) error between soil moisture measured by GPR and TDR was 0.038 m3·m-3. For the radiometer, the rms error decreased from 0.062 (horizontal polarization) and 0.054 (vertical polarization) to 0.020 m3·m-3(both polarizations) after accounting for roughness using an empirical model that required calibration with reference TDR measurements. Monte Carlo simulations showed that around 20% of the reference data were required to obtain a good roughness calibration for the entire field. It was concluded that relatively accurate measurements were possible with both methods, although accounting for surface roughness was essential for radiometry.

We used advanced ground-penetrating radar (GPR) inversion techniques for detecting landmines in laboratory conditions. The radar data were acquired with a calibrated vector network analyzer combined with an off-ground monostatic horn antenna, thereby setting up a stepped-frequency continuous-wave radar. Major antenna effects and interactions with the soil and targets were filtered out using frequency-dependent complex antenna transfer functions. The proposed strategy first exploits inversion approaches that are able to give an accurate characterization of the antenna-soil interaction and a reliable estimate of the soil permittivity. The outcomes of this first phase are at the basis of the application of a microwave tomographic approach based on the Born approximation to achieve the imaging of the subsurface. The algorithms were applied for imaging three landmines of different sizes and buried at different depths in sand. Although the radar system was off the ground, the results showed that it was possible to reconstruct all mines, including a shallow plastic mine as small as 5.6 cm in diameter. This last mine was invisible in the raw radar data, and the use of common GPR imaging techniques did not lead to satisfactory results. The proposed integrated method shows great promise for shallow subsurface imaging in a demining context, particularly because it automatically provides accurate information on the shallow soil dielectric permittivity.

We analyzed the effect of shallow thin layers on the estimation of soil surface water content using full-waveform inversion of off-ground ground penetrating radar (GPR) data. Strong dielectric contrasts are expected to occur under fast wetting or drying weather conditions, thereby leading to constructive and destructive interferences with respect to surface reflection. First, synthetic GPR data were generated and subsequently inverted considering different thin-layer model configurations. The resulting inversion errors when neglecting the thin layer were quantified, and then, the possibility to reconstruct these layers was investigated. Second, laboratory experiments reproducing some of the numerical experiment configurations were conducted to assess the stability of the inverse solution with respect to actual measurement and modeling errors. Results showed that neglecting shallow thin layers may lead to significant errors on the estimation of soil surface water content(¿¿>0.03 m3/m3), depending on the contrast. Accounting for these layers in the inversion process strongly improved the results, although some optimization issues were encountered. In the laboratory, the proposed full-waveform method permitted to reconstruct thin layers with a high resolution up to 2 cm and to retrieve the soil surface water content with an rmse less than 0.02 m3/m3, owing to the full-waveform inverse modeling. These results suggest that the proposed GPR approach is promising for field-scale mapping of soil surface water content of nondispersive soils with low electrical conductivity and for instances when soil layering is encountered.

The detection of antipersonnel landmines using ground-penetrating radar (GPR) is particularly hindered by the predominant soil surface and antenna reflections. In this paper, we propose a novel approach to filter out these effects from 2-D off-ground monostatic GPR data by adapting and combining the radar antenna subsurface model of Lambotwith phase-shift migration. First, the antenna multiple reflections originating from the antenna itself and from the interaction between the antenna and the ground are removed using linear transfer functions. Second, a simulated Green's function accounting for the surface reflection is subtracted. The Green's function is derived from the estimated soil surface dielectric permittivity using full-wave inversion of the radar signal for a measurement taken in a local landmine-free area. Third, off-ground phase-shift migration is performed on the 2-D data to filter the effect of the antenna radiation pattern. We validate the approach in laboratory conditions for four differently detectable landmines embedded in a sandy soil. Compared to traditional background subtraction, this new filtering method permits a better differentiation of the landmine and estimation of its depth and geometrical properties. This is particularly beneficial for the detection of landmines in low-contrast conditions.

The possibility to estimate accurately the subsurface electric properties from ground-penetrating radar (GPR) signals using inverse modeling is obstructed by the appropriateness of the forward model describing the GPR subsurface system. In this paper, we improved the recently developed approach of Lambot et al. whose success relies on a stepped-frequency continuous-wave (SFCW) radar combined with an off-ground monostatic transverse electromagnetic horn antenna. This radar configuration enables realistic and efficient forward modeling. We included in the initial model: 1) the multiple reflections occurring between the antenna and the soil surface using a positive feedback loop in the antenna block diagram and 2) the frequency dependence of the electric properties using a local linear approximation of the Debye model. The model was validated in laboratory conditions on a tank filled with a two-layered sand subject to different water contents. Results showed remarkable agreement between the measured and modeled Green's functions. Model inversion for the dielectric permittivity further demonstrated the accuracy of the method. Inversion for the electric conductivity led to less satisfactory results. However, a sensitivity analysis demonstrated the good stability properties of the inverse solution and put forward the necessity to reduce the remaining clutter by a factor 10. This may partly be achieved through a better characterization of the antenna transfer functions and by performing measurements in an environment without close extraneous scatterers.