Terrorist attacks pose a severe threat to global public security, among which pre-embedding offensive weapons in walls is an important attack type. Therefore, obtaining information about buried objects is crucial for addressing potential threats. Ground-penetrating radar (GPR), as an established non-intrusive detection method, is capable of effectively acquiring the three-dimensional (3D) information of buried objects. However, traditional GPR imaging methods often fail to complete the 3D reconstruction task for targets under conditions of low signal-to-noise ratio (SNR). Moreover, the large computational load limits its application in real-time detection scenarios. To address these challenges, this paper proposes a robust and lightweight 3D reconstruction method for buried targets, specifically for pistols, eavesdropping devices, and explosives, etc. The method first utilizes the 3D Fourier transform of the wave equation to analyze the target wavefield information from 3D C-scan data, which can obtain a preliminary visualization of the target, then employs a projection mechanism to process the sparse energy-focused volume from multiple views. Finally, a multi-view reconstruction algorithm is used to reconstruct the 3D voxel model of the target. Compared to existing methods, our approach reduces parameters by 98.9% to 48.59M versus 3D U-Net, FLOPs by 79.7% versus Kirchhoff migration, and achieves 0.575-second inference with a 0.829 Dice coefficient under optimal conditions. It maintains robust performance down to -5 dB SNR, with monotonic improvements in MSE, MAE, and Dice as SNR increases. Experimental results indicate that the approach facilitates rapid and accurate retrieval of information on concealed dangerous objects in resource-limited settings.

The hyperbola detection in the ground-penetrating radar (GPR) data is of real significance for subsurface object localization. However, present detection methods in GPR cannot process B-scan data with arbitrary length. In this letter, a hyperbola detection method based on the state-space model (SSM) for arbitrarily long GPR B-scan is proposed. The proposed method consists of three parts: a time dimension encoder based on the ResNet block, an SSM module, and a time dimension decoder based on the transposed convolution. First, the time dimension encoder extracts high-dimensional time dimension signal features from each scan channel of B-scan data. Then, the SSM module extracts bidirectional correlation features along the survey line from the feature maps obtained in the first part, thereby obtaining feature maps containing hyperbolic target features. Finally, the time dimension decoder decodes the features obtained in the second part and outputs the probability map representing the target hyperbolic vertex region. Based on the probability map, the hyperbola detection and localization results can be obtained. The effectiveness of the proposed method is verified by both simulation and field experiments using GPR B-scan data. In addition, the experimental results show that the proposed method can achieve the AP with 96.9%. The code is available athttps://github.com/yimaxwell/state-space-model-for-GPR-detection.git.

Pipeline and void recognition are two key tasks in the underground monitoring of urban roads. Ground penetrating radar (GPR), as an effective geophysical method, plays an important role in this area. With the increasing amount of GPR data, automatic recognition has become a research hotspot. However, existing automated recognition methods still suffer from low accuracy or high dependence on datasets. In this paper, a non-learning method for both pipeline and void recognition is proposed. In this method, the B-scan is preprocessed by removing the direct coupled wave and multiple echoes.Then, the binarized image is converted to sparse image using non-zero interval sparse coding (NISE). Next, the slope distribution of sparse images is extracted using column offset coding (COE). And clustering is carried out according to the corresponding relationship between the image slope and its original position. Finally, the decision is made according to the slope distribution characteristics of each cluster. The proposed method was tested on both simulated and field data. Experimental results show that the method not only has the advantage of being training-free but also exhibits excellent recognition accuracy.

In this article, we present a layered cross correlation back-projection (LCBP) method based on ray-theory for electromagnetic (EM) imaging in stratified medium. To argue the challenge of resolving time delay in the stratified medium in traditional time-domain methods, we first construct a ray propagation model, then calculate the time delay at any point on the path and solve the sparse time delay at the deep position by linear interpolation, Finally, a clutter suppression method is used to reduce interference through the cross correlation of signals between different subapertures. LCBP achieves fast imaging in a stratified medium environment whereas retaining the advantages of the BP algorithm in the focusing ability. In order to verify the performance of the proposed method, the simulation and experimental data are carried out with comparisons of the phase shift migration (PSM) imaging method and layered range migration (LRM), and the proposed method exhibits excellent imaging quality.

Layered parameter estimation represents a crucial application of ground penetrating radar (GPR), playing a pivotal role in reconstructing the internal structures of media. In situations where adjacent media are similar and multiple reflected waves are present, conventional methods face substantial challenges in detecting weak echoes and accurately extracting time delays. To achieve precise estimation of layered media parameters in the presence of similar materials and multiple reflected waves, this paper presents a modified Orthogonal Matching Pursuit (OMP) parameter estimation method. This method extracts the correct time delays by eliminating multiple reflected waves and integrates the constructed generalized reflection coefficients to recover the signal. Subsequently, a genetic algorithm is utilized to optimize the constructed objective function, enabling the precise estimation of the thicknesses and permittivities of adjacent layers with similar media. This method is particularly applicable to scenarios involving two or more subsurface layers with nearly identical permittivities. Finally, numerical and real experiments have been conducted to validate the accuracy and effectiveness of the proposed method.

Due to the electromagnetic propagation loss and environmental interference, the deep ground penetrating radar (GPR) image is poor for further interpretation. Many image enhancement methods including resolution improvement and clutter removal have been widely studied to improve the GPR image quality. In this letter, a deep GPR image enhancement method is proposed to generate clear high-resolution images by the diffusion model (DM). In order to ensure the model is equipped with abilities of resolution enhancement and declutter simultaneously, the low-frequency image inserted with clutter performances as prior knowledge input network to fit the Gaussian distribution clutter added in the forward process of the DM. The method has already been tested by simulation and field experiments. Compared with the classical methods of declutter only and improving the resolution only, our method achieves the best results by synthesizing entropy (EN), peak signal-to-clutter ratio (PSCR), and structural similarity (SSIM), which proves the effectiveness of resolution enhancement and clutter removal in the deep GPR image.

The recognition of underground pipelines is an important in urban areas. As an efficient and non-destructive recognition method, ground penetrating radar (GPR) has been increasingly applied in the recognition of underground pipelines. With the growing volume of GPR data, there is an urgent need for automatic recognition. However, due to the complexity of the subsurface environment, existing automatic recognition methods still have drawbacks such as low accuracy, poor robustness, and the requirement for large training datasets. An underground pipeline recognition model for GPR that combines scale-invariant feature transform (SIFT) and support vector machine (SVM) is proposed in this article. The model is based on the fact that there are scale-invariant keypoints at the tops of hyperbolas. First, SIFT is used to identify scale-invariant keypoints in the image. These keypoints undergo symmetry assessment and feature enhancement. Subsequently, SVM is employed to filter out the keypoints located at the tops of the hyperbolas. Finally, keypoints located on the same hyperbola are clustered to obtain the recognition results. The model improves the original SIFT method by modifying the calculation of the blur coefficients in the Gaussian pyramid layers. It also employs manually designed feature enhancement methods when constructing the feature descriptors. Additionally, we have introduced methods such as symmetry judgment to further enhance the model’s accuracy. The results indicate that the proposed method exhibits superior recognition performance for the field data even with very limited training samples.

Layered medium parameter inversion is a crucial technique in ground-penetrating radar (GPR) data processing and has wide application in civil engineering and geological exploration. In response to the issues of high computational complexity and low accuracy associated with existing methods, a novel layered medium parameter inversion approach is proposed, comprising the deconvolution autoencoder and the parameter inversion network. First, the deconvolution autoencoder is introduced to solve the pulse response of layered medium systems in an unsupervised manner, which enhances the computational efficiency of deconvolution and decouples the data acquisition system from the supervised model. Subsequently, a parameter inversion network, including a self-attention module and a residual multilayer perceptron (MLP), is proposed to address the challenge posed by the excessively sparse pulse responses. The self-attention module calculates the autocorrelation of the pulse sequence, providing temporal delay information between pulses and reducing the sparsity of the pulse response to facilitate feature extraction. Meanwhile, the residual MLP, known for its low information loss and adaptability to different output dimensions, is employed for model-based and pixel-based inversions in situations with and without prior knowledge of the layer number, respectively. Finally, simulated and measured datasets are constructed to comprehensively train and evaluate the proposed method. The results demonstrate that the proposed method exhibits better performance of inversion accuracy, computational efficiency, robustness, generalization capability, and noise resistance. In addition, it remains applicable even when prior knowledge of the layer number is unknown.

Media parameter inversion is an important issue in the field of electromagnetic propagation widely applied to multilayered scenarios in ground penetrating radar (GPR). However, due to unsolvable refraction points in multilayered conditions, computing burden in iterative optimization, and unreachable global optima in multi-objective optimization, the existing inversion methods cannot effectively realize the multilayered inversion. For the problem, demanding for high precision and high efficiency in multilayered applications, a multilayered media parameter inversion based on reflected trajectory fitting is proposed. The method adopts a layer-by-layer inversion mechanism with a novel refraction approximation. At each layer inversion, it utilizes the reflected travel equation of the first A-scan and the reflected trajectory fitting by least square linear regression to draw two relation curves whose intersection position determines layer thickness and permittivity. Finally, through simulations and experiments, the method shows its accuracy and effectiveness in the multilayered structure.

The performance of ground penetrating radar (GPR) target detection is seriously affected by the clutter. In this letter, an effective GPR clutter removal method is proposed based on low-rank and sparse decomposition with total variation regularization (LRSD-TVR). In the proposed method, a total variation (TV) regularization of sparse matrix is introduced to further remove the remaining clutter and to obtain a clearer target image. An iterative approach based on alternating direction method of multipliers (ADMM) is developed to solve the optimization problem of LRSD-TVR. In each iteration, the low rank component which corresponds to the clutter is computed by singular value decomposition (SVD) thresholding. Besides, the sparse component corresponding to the target is obtained by solving the sub-optimization problem reformulated in terms of TV component. The effectiveness of proposed method is verified by both numerical simulations and field experiments.

Ground-penetrating radar (GPR) is an effective detection tool for multilayered structures such as road detection and underground exploration. However, traditional inversion algorithm in GPR cannot achieve layered media parameter inversion accurately due to neglect of multiple reflections and complicated refraction effect. To address this issue, an inversion algorithm for layered media parameters based on common middle point (CMP) model and pattern search (PS) method is proposed in this letter. The proposed method combines the wave propagation equation and Snell’s law to build the cost functions in CMP model, and takes the layered media parameters as the optimization variables. Both numerical and experiment results indicate that inversion efficiency and accuracy can be improved. The novel inversion algorithm saves at least a quarter of time to control relative errors within 1% and 6% in simulation and laboratory experiment, respectively. The field test in asphalt highway shows its potential benefits in the field of multilayered structure.

In order to solve the problems of reduced accuracy and prolonging convergence time of through-the-wall radar (TWR) human motion due to wall attenuation, multipath effect, and system interference, we propose a multi-link auto-encoding neural network (TWR-MCAE) data augmentation method. Specifically, the TWR-MCAE algorithm is jointly constructed by a singular value decomposition based data preprocessing module, an improved coordinate attention module, a compressed sensing learnable iterative shrinkage threshold reconstruction algorithm (LISTA) module, and an adaptive weight module. The data preprocessing module achieves wall clutter, human motion features, and noise subspaces separation. The improved coordinate attention module achieves clutter and noise suppression. The LISTA module achieves human motion feature enhancement. The adaptive weight module learns the weights and fuses the three subspaces. The TWR-MCAE can suppress the low rank characteristics of wall clutter, and enhance the sparsity characteristics in human motion at the same time. It can be linked before the classification step to improve the feature extraction capability without adding other prior knowledge or recollecting more data. Experiments show that the proposed algorithm gets a better peak signal-to-noise ratio (PSNR), which increases the recognition accuracy and speeds up the training process of the back-end classifers.

To better reconstruct underground targets based on ground-penetrating radar (GPR) data, this paper proposes a joint physics and data driven full-waveform inversion (PDD-FWI) scheme. This scheme combines a physics-based non-iterative approach and a data-driven deep neural network (DNN) to reconstruct target location, shape and permittivity accurately. Firstly, the normalized range migration algorithm (RMA) is introduced to extract the target contour and location information, which not only improves the robustness of the proposed scheme, but also ensures adaptability to different GPR equipment. Then, the GPR dielectric target reconstruction network (GPRDtrNet) is developed based on the improved U-net structure, including reducing network layers and adding multi-scale additive spatial attention gates and skip-connection structures. Compared with previous DNN-based reconstruction methods, GPRDtrNet has the advantages of small data requirement, high accuracy, strong generalization and noise tolerance. Finally, the simulated and real dataset containing kinds of targets is constructed to train and test GPRDtrNet. The results show that the proposed method can reconstruct underground dielectric targets accurately with high robustness and noise tolerance.

A joint petrophysical and structural inversion method for electromagnetic (EM) and seismic data based on the volume integral equation (VIE) is proposed in this paper. In the forward EM problem, only the contrast of conductivity is solved by the electric field integral equation method. However, in the forward seismic problem, both the contrasts of velocity and mass density are solved by the combined field VIE method. Both forward solvers are accelerated by the fast Fourier transform. In the inversion problem, by using the petrophysical equations about the porosity and saturation and applying the chain rule, we fuse the EM and seismic data and construct the joint petrophysical inversion equations, which can be solved by the variational Born iteration method. Then, in order to further enhance the reconstructed results of the joint petrophysical inversion, we enforce the structural similarity constraint between porosity and water saturation and add the cross-gradient function to the joint petrophysical inversion cost function. Two typical geophysical models based on the remote sensing measurement are used to validate the proposed methods. One is the cross-well model, and the other is the marine surface exploration model. The advantage of the joint inversion compared with the separate inversion is evaluated based on the resolution and the data misfits of the reconstructed profiles as well as the antinoise ability.

An efficient 2-D joint full-waveform inversion method for electromagnetic and seismic data in a layered medium background is developed. The joint inversion method based on the integral equation (IE) method is first proposed in this paper. In forward computation, the IE method is employed, which usually has smaller discretized computation domain and less cumulative error compared with the finite-difference method. In addition, fast Fourier transform is used to accelerate the convolution between Green's functions and induced sources due to the shift invariance property of the layered Green's functions in the horizontal direction. In the inversion model, the cross-gradient function is incorporated into the cost function of the separate inversion to enforce the structure similarity between electric conductivity and seismic-wave velocity. We use the improved variational Born iteration method and two different iteration strategies to minimize the cost function and reconstruct the contrasts. Several typical models in geophysical applications are used to validate our joint inversion method, and the numerical simulation results show that joint inversion can improve the inversion results when compared with those from the separate inversion.