Ground Penetrating Radar B-scan Image Research Articles

Experimental validation of the horizontal resolution improvement by ultra-wideband metasurfaces for GPR systems

Achieving high horizontal resolution is crucial for general non-destructive testing (NDT) applications, and in the domain of ground penetrating radar (GPR), it can be more challenging due to inefficient transmissions of electromagnetic waves into and through complex material under test (MUT). In this study, we showcase the validation results of our recently designed ultra-wideband metasurface (UWM) in improving the horizontal resolution for GPR B-Scan images. This UWM has an ultra-wide working frequency band (100%) from 1 to 3 GHz, in which high-frequency GPR signals can be enhanced and transmitted into MUT. Our fourteen-day consecutive GPR experiments demonstrate that two buried pipes with an edge-to-edge spacing of 8 cm that are hard to distinguish by traditional GPR surveys can be visually identified by depositing the UWM atop the MUT. The underlying mechanism is the sustained and improved transmission of high-frequency signals into the MUT, enabled by the removal of transmission coupling loss by the UWM at the air–MUT interface and the resulting enhanced transmission of more high-frequency components. Our simulations also provide quantitative analysis of such enhanced behavior with a nominal transmittance increase of 30% ∼50%. We believe the horizontal resolution improvement enabled by UWM will open a new corridor for high-resolution GPR system design.

REN-GAN: Generative adversarial network-driven rebar clutter elimination network in GPR image for tunnel defect identification

In ground penetrating radar (GPR) images, the signals from reinforcing steel bars (rebars) can significantly obscure the signals of void defects beneath them, which hinders defect detection in tunnel linings. This study introduces a novel, end-to-end unsupervised deep learning-based network that is trained on unpaired GPR images. The primary objective is to suppress the rebar signal and enhance the accuracy of void defect identification in recorded GPR B-Scan images. The network incorporates an autoencoder generator with feature attention and feature mix-up modules, designed to recover textural information and generate improved images devoid of rebar interference. Crucially, this network leverages a strategy of comparing images in feature space, rather than pixel space. The features extracted by the autoencoder are constrained by cycle perceptual consistency loss, derived from a pretrained VGG19 network, to ensure fidelity and perceptual accuracy. Furthermore, this study introduces an enhanced Faster-RCNN network, specifically improved through a more robust feature extraction module, for detecting void defects beneath rebars. The effectiveness of the proposed method has been thoroughly validated using both synthetic and real-world data sets. The results indicate that our method not only surpasses other unsupervised techniques in performance but also effectively diminishes rebar signals and accurately reconstructs and identify void defect signals.

Underground Diagnosis Based on GPR and Learning in the Model Space.

Ground Penetrating Radar (GPR) has been widely used in pipeline detection and underground diagnosis. In practical applications, the characteristics of the GPR data of the detected area and the likely underground anomalous structures could be rarely acknowledged before fully analyzing the obtained GPR data, causing challenges to identify the underground structures or anomalies automatically. In this article, a GPR B-scan image diagnosis method based on learning in the model space is proposed. The idea of learning in the model space is to use models fitted on parts of data as more stable and parsimonious representations of the data. For the GPR image, 2-Direction Echo State Network (2D-ESN) is proposed to fit the image segments through the next item prediction. By building the connections between the points on the image in both the horizontal and vertical directions, the 2D-ESN regards the GPR image segment as a whole and could effectively capture the dynamic characteristics of the GPR image. And then, semi-supervised and supervised learning methods could be further implemented on the 2D-ESN models for underground diagnosis. Experiments on real-world datasets are conducted, and the results demonstrate the effectiveness of the proposed model.

A novel classification method for GPR B-scan images based on weak-shot learning

Ground Penetrating Radar (GPR) is extensively employed for underground target detection, and the advancements in this domain have been hastened by the utilization of deep learning methodologies. Despite advancements in deep learning, practical engineering applications often face challenges in meeting the demanding requirements of large-scale data and high data quality. In this study, based on cross-domain transfer learning, we proposed a method for underground target classification of weak labeled GPR data. It combined strongly annotated source domain data with weakly annotated target domain data through adversarial learning to leverage the prior knowledge from existing large datasets and addressed the classification problem of small-sample GPR data. By employing adversarial learning to integrate strongly labeled source domain data with weakly labeled target domain data, we leverage prior knowledge from existing large datasets, thereby enhancing classification accuracy on a small GPR dataset. This approach reduces the requirements for annotation quality and sample quantity in the GPR B-scan dataset of underground targets. Additionally, it exhibits robustness to label noise, making it applicable to real-world GPR data classification tasks. Compared to some well-known networks, the proposed model achieved the best classification performance in multiclass classification problems, with an accuracy of 87.62% and an F1macro score of 87.60% on the test set. These results represent an improvement of 17.14 and 17.20 percentage points, respectively, over the best-performing classical classification networks. Additionally, the proposed model's reliability in cross-domain transfer learning, even with limited data, was substantiated through a comparison with other few-shot learning networks.

Deep Learning Inversion of Multi-processed GPR Data to Characterize Subsurface Defects

Ground penetrating radar (GPR) is highly recommended for investing defects in concrete structures, yet interpreting GPR data is hindered by the fact that the images primarily display reflected electromagnetic waves rather than the actual shapes of the defects. In this paper, we introduce a deep learning network that aims to convert GPR B-scan images into permittivity maps of subsurface structures. This network takes three types of data as inputs and outputs the distribution of permittivity. These three types of data are the original data, the data obtained after applying migration to the original data, and the data obtained after processing the original data using a neural network. Each type of data is processed using different methods and contains distinct information, which is collectively used for inverting data. In this study, we validate the effectiveness of the proposed network by the finite-difference time-domain (FDTD) forward modeling data.

Automatic recognition of defects behind railway tunnel linings in GPR images using transfer learning

Ground-penetrating radar (GPR) is widely used for non-destructive detection of hidden defects behind railway tunnel linings, including non-compactness, air- and water-filled voids, fractures, and insufficient lining thickness. However, manual interpreting a substantial volume of GPR B-scan images is time-consuming, and the results heavily depend on the professional knowledge of the practitioner and the available prior information. This paper presents an automatic recognition algorithm for detecting hidden defects, as well as estimating lining thickness in GPR B-scan images of railway tunnels using a You Only Look Once for Panoptic driving perception network (YOLOP) model. The process mainly contains three steps. Firstly, a real GPR dataset of underground objects is employed for transfer learning to enhance the robustness of the prediction model. Secondly, a dataset containing 2,245 real GPR B-scan images of railway tunnel linings is created. Thirdly, the established dataset is used to retrain the YOLOP model and assess the prediction accuracy for identifying hidden defects and lining thickness in GPR images. Field experiment results demonstrate that the trained hidden defects recognition algorithm achieves an average accuracy of 80.6%, with a detection time of only 0.02 s for a GPR image sizing of 1024 × 512. Therefore, it is concluded that the proposed algorithm is efficient and reliable for automatic recognition of hidden defects behind railway tunnel linings.

Depth and Radius Joint Estimation for Underground Pipeline Using GPR

In this paper, a deep learning-based methodology is proposed to estimate depth and radius for underground pipeline using Ground penetrating radar(GPR) images. The proposed methodology conducts the depth and radius joint estimation based on GPR B-scan image preprocessing. During the preprocessing, the autoencoder structure network is designed to binarize the GPR image. Besides, redundant detection boxes of hyperbola detection are also eliminated. For pipeline depth estimation, the parabola model is used instead of the hyperbola model to obtain the symmetry axis of the pipeline GRP image more efficiently. The consecutive a-scan data at extracted symmetry axis and its vicinity are spliced and sent to the Long Short-Term Memory Network for depth estimation. For pipeline radius estimation, the image features and depth information are concatenated. A pre-trained convolutional neural network is designed to extract the GPR image features. Meanwhile, a 2d-Mapping network is proposed to encode the 1-dimensional depth information into 2-dimensional feature matrix. Improving the dimension of depth information rather than flattening the GPR image features (dimension reduction) can preserve the spatial information of image features. Besides, the depth information can be concatenated with the image features more flexibly. In all, the proposed methodology can accurately estimate the depth and radius of underground pipeline. To verify the effectiveness of the proposed methodology, the gprMax is used to generate the pipeline GPR images in inhomogeneous soil and create the dataset. The experiments include detailed comparison and analysis, illustrating that the proposed methodology is feasible and effective.

A Modular Method for GPR Hyperbolic Feature Detection and Quantitative Parameter Inversion of Underground Pipelines

Ground penetrating radar (GPR) is widely used to inspect underground pipelines because it is non-destructive. When the scan line of GPR is perpendicular to the pipe, it will exhibit hyperbolic features in GPR B-scan images, which have no intuitive relationship with the geometric and physical parameters of the pipeline, making the interpretation of GPR images difficult. This paper proposes a modular detection and quantitative inversion method for the hyperbolic features in GPR B-scan images, which is divided into two steps. In the first step, the YOLOv7 object detection network is used to automatically detect the hyperbolic features in GPR images. In the second step, a two-stage curve fitting method is proposed based on the characteristics of the detection model. It uses a few key point annotations of the hyperbolic pattern and some parameters of the GPR system to quantitatively invert the depth and radius of pipes. Using the same hardware and data set, YOLOv7 achieves an 11.1% improvement in detection accuracy and an 18.2% improvement in speed compared to YOLOv5. The relative errors of the proposed method for the depth and radius of the synthetic data in homogeneous media are 0.6% and 4.4%, respectively, and 4.8% and 15% in non-homogeneous media. The relative error of the depth inversion of the measured data TU1208 is less than 10%. The results show that the method can effectively invert the depth and radius of underground pipelines and reduce the difficulty of GPR data interpretation.

Unsupervised learning method for rebar signal suppression and defect signal reconstruction and detection in ground penetrating radar images

The signals of reinforcing steel bars (rebars) in ground penetrating radar (GPR) images may mask defect signals underneath them, negatively affecting defect detection in reinforced concrete. This study proposes an unsupervised learning-based method trained using unpaired GPR images to suppress the rebar signal and reconstruct the defect signal on the recorded GPR B-Scan images. In particular, four contrastive feature encoders and two similarity feature encoders are designed in the network. The extracted features of contrastive feature encoders and similarity feature encoders are constrained by contrastive and similarity loss functions, respectively. The proposed method was validated using data from three scenarios: synthetic data, sandbox experimental data, and data collected from reinforced concrete in field. The results showed that the proposed method outperforms other unsupervised methods, and it can effectively suppress the rebar signals and accurately reconstruct the defect signals. Moreover, the identification accuracy of defect underneath rebars can be improved significantly.

Simulation of GPR B-Scan Data Based on Dense Generative Adversarial Network

Urban subsurface infrastructures, e.g., pipelines and roads, are aging with the expansion of modern cities. Benefiting from the capability of nondestructive detection, ground penetrating radar (GPR) has been widely applied to underground objects or disasters detection, and GPR B-scan images are employed by manual interpretation. While, this way of high subjectivity and uncertainty inevitably results in failure of detection. Meanwhile, the shortage of labelled images greatly impedes the automatization and intelligentization of underground disaster detection based on GPR. Many data simulation techniques, e.g., forward modelling, were used to augment images for training; however, the generated forward images were not similar enough to the real B-scan data, which makes recognition a challenging task. To address this problem, we proposed a novel B-scan image simulation method based on generative adversarial network to generate synthetic images for training detection networks. Our network utilizes DenseNet as the backbone network of generator to extract image features, and a weighted total variation regularization term to regularize the loss function of the network. The comparison and ablation experiments verified that our network could generate simulation images with high similarity to real GPR B-scan images. We believe that this work contributes to the intelligent processing and analysis of GPR data, and improves the efficiency of underground disaster detection.

Signal Calibration and Noise Reduction for Software-Defined Radio-Based GPR

This article studies signal calibration, frequency response estimation, and noise reduction for a software-defined radio (SDRadio)-based ground-penetrating radar (GPR). The radar signal generation utilizes the stepped frequency continuous wave (SFCW) method. The signal calibration takes advantage of the dual-channel structure of the SDRadio and solves the incoherent phase problem caused by phase mismatch between the local oscillators (LOs) in the transmitter and the receiver. The frequency response estimation involves two steps: First, the calibration measures the system’s intrinsic frequency response. Then it measures the frequency response of the target under the ground surface. The noise effect at the receiver is reduced using the ridge regression method. For performance evaluation, outdoor and indoor experiments are conducted, which show that the GPR A-scan waveform and B-scan image qualities can be effectively improved. Compared with the traditional method, the signal-to-clutter ratio (SCR) of the A-scan waveform is improved by around 7 dB and that of the B-scan image is improved by around 13 dB.

Robotic Inspection of Underground Utilities for Construction Survey Using a Ground Penetrating Radar

Ground penetrating radar (GPR) is a very useful nondestructive evaluation (NDE) device for locating and mapping underground assets prior to digging and trenching efforts in construction. This paper presents a novel robotic system to automate the GPR data-collection process, localize underground utilities, and interpret and reconstruct the underground objects for better visualization, allowing regular nonprofessional users to understand the survey results. This system is composed of three modules: (1) an omnidirectional robotic data-collection platform that carries a RGB-D camera with inertial measurement unit (IMU) and a GPR antenna to perform automatic GPR data collection and tag each GPR measurement with visual positioning information at every sampling step, (2) a learning-based migration module to interpret the raw GPR B-scan image into a two-dimensional (2D) cross-section model of objects, and (3) a three-dimensional (3D) reconstruction module, i.e., 30.0% GPRNet, to generate underground utility model represented as fine 3D point cloud. Comparative studies were performed on synthetic data and field GPR raw data with various incompleteness and noise. Experimental results demonstrated that our proposed method achieves a higher GPR imaging accuracy in mean intersection over union (IoU) than the conventional back-projection (BP) migration approach, and 6.9%−7.2% less loss in Chamfer distance (CD) than point cloud model reconstruction baseline methods. The GPR-based robotic inspection provides an effective tool for civil engineers to detect and survey underground utilities before construction.

Using dictionary learning for clutter reduction in GPR B-scan images

ABSTRACT Ground-penetrating radar (GPR) has been widely used to detect subsurface objects. However, the target reflection experiences interference owing to clutter, such as direct coupling, interface reflection, and other undesired reflections. Reducing such clutter is a valuable processing step to improve detection accuracy. This letter proposes a GPR clutter reduction algorithm based on dictionary learning (DL). DL learns an adaptive dictionary that can sparsely represent the GPR B-scan image and enable clutter suppression. The learned dictionary atoms are divided into target atoms and clutter atoms, the target component and the clutter component can be reconstructed using these two types of dictionary atoms, respectively. The performance of the proposed algorithm is validated using both simulation data and real GPR data. The results of the visual evaluation and quantitative analysis demonstrate that the proposed algorithm outperforms the existing clutter reduction algorithms.

Automatic recognition of pavement cracks from combined GPR B-scan and C-scan images using multiscale feature fusion deep neural networks

Pavement crack detection is critical for transportation infrastructure assessment using ground penetrating radar (GPR). This paper describes a YOLOv3 model with four-scale detection layers (FDL) to detect combined B-scan and C-scan GPR images subject to poor detection effects and a high missed detection rate of small crack feature sizes. Multiscale fusion structures, efficient intersection over union (EIoU) loss function, K-means++ clustering, and hyperparameter optimization were used in this proposed model to further improve detection performance. Results indicated that the F1 score and mAP of the YOLOv3-FDL model reached 88.1% and 87.8% and had an 8.8% and 7.5% improvement on the GPR dataset of concealed cracks, respectively, compared with the YOLOv3 model. This illustrated that this model solved the problem of missed crack detection to some extent. Future studies can take these results further, especially the three-dimensional feature analysis of pavement cracks.

Automatic recognition and localization of underground pipelines in GPR B-scans using a deep learning model

Ground penetrating radar (GPR) is a popular non-destructive method for detecting and locating underground pipelines. However, manual interpretation of a large number of GPR B-scan images is time-consuming, and the results highly relies on the practitioner’s experience and the priori information at hands. An automatic GPR method for recognition and localization of underground pipelines is proposed based on a deep learning model in the paper. Firstly, a dataset containing 3,824 real GPR B-scans of pipelines is established. Secondly, a You Only Look Once version 3 (YOLOv3) model is trained to recognize the regions of the underground pipelines in a GPR image. Thirdly, the hyperbolic response of a pipeline is focused by migration, and transformed into a binary image by an iterative thresholding method. Finally, the apex of the hyperbola is employed to estimate both the horizontal position and the buried depth of the pipeline. Field experiments validated that the absolute errors of the estimated depths are less than 0.04 m and the average relative error is lower than 4 %. It is demonstrated that the proposed method is automatic, high-speed, and reliable for recognition and localization of underground pipelines in urban area.

Investigating Transfer Learning Performances of Deep Learning Models for Classification of GPR B-Scan Images

Recent advances in deep learning models have made them the state-of-the art method for image classification. Due to this success, they have been applied to many areas, such as satellite image processing, medical image interpretation, video processing, etc. Recently, deep learning models have been utilized for processing Ground Penetrating Radar (GPR) data as well. However, studies general focus on building new Convolutional Neural Network (CNN) models instead of utilizing baseline ones. This paper investigates the usefulness of existing baseline CNN models for classifying GPR B-scan images and aims to determine how well pre-trained models perform. To that end, a real bridge deck GPR data, DECKGPRHv1.0 dataset was used to evaluate the transfer learning performances of various CNN models. Different variants of the models in terms of varying depths and number of parameters were also considered and evaluated in a comparative manner. Although it is an older model, ResNet achieved the best results with 0.998 accuracy. The experimental results showed that there is generally a direct correlation between the simplicity of the model and its success. Overall, it is concluded that near perfect results are possible by just adapting pre-trained models to the problem without fine-tuning.

A Multi-Path Encoder Network for GPR Data Inversion to Improve Defect Detection in Reinforced Concrete

Ground penetrating radar (GPR) has been extensively used in the routine inspection of reinforced concrete structures. However, the signatures in GPR images are reflected electromagnetic waves rather than their actual shapes. The interpretation of GPR data is a mandatory but time- and labor-consuming task. Furthermore, the rebars in the near-surface of concrete cause clutter in the GPR images, which hinders the interpretation of GPR data. This work presents a deep learning network to invert GPR B-scan images to permittivity maps of subsurface structures. The proposed network has a multi-path encoder which enables the network to leverage three kinds of GPR data: the original, migrated, and encoder–decoder-processed GPR data. Each type of processing method is designed to serve a different purpose: the original GPR images retain all the waveforms; the migration method intensifies the vertices of the subsurface anomalies; the encoder–decoder network suppresses rebar clutter and enhances the visibility of the defect echoes. The outputs of three processing methods are jointly used to interpret GPR B-scan images. We demonstrated the superiority of the proposed network by comparing it with a network with a single-path encoder. We also validated the proposed network with synthetic and experimental GPR data. The results indicate that the proposed network effectively reconstructs the defects in the reinforced concrete.

Training data augmentations for improving hyperbola recognition in ground penetrating radar B-scan image for tree roots detection

Improving the detection accuracy of hyperbola in B-scan images has been a considerable challenge for ground penetrating radar (GPR) to detect tree roots. In this paper, a method for data enhancement and target detection, both based on deep learning was proposed to identify hyperbolas in GPR B-scan images. First, the authors used a cyclic consistent adversarial network (CycleGAN) to augment the original data. In this procedure, the hyperbolic features of the images were preserved and created a wider variety of training samples. Then, the authors could apply the enhanced dataset to the YOLOv5 detection model to evaluate the effectiveness of their method. Meanwhile, the detection effects of Yolov3, Yolov5, Faster R-CNN, and CenterNet detection models on the enhanced dataset were compared. The results showed that applying the enhanced dataset to the Yolov5 detection model exhibited better detection accuracy compared to other combinations of datasets and detection models. The authors demonstrate that the proposed method increases data diversity and the number of samples, improving the precision and recall of hyperbolic curves. These results provide a new method for tree root localization with important effects.

Simultaneous tunnel defects and lining thickness identification based on multi-tasks deep neural network from ground penetrating radar images

The overall assessment of tunnel lining, including shapes, categories, and depths of tunnel internal defects as well as the thickness of tunnel linings is vital to the safe operation of tunnels. We proposed a method comprising a multi-task deep neural network and curve fitting post-processing operation for simultaneously identifying the shapes, categories, and depths of tunnel defects as well as lining thicknesses from ground penetrating radar (GPR) images. The multi-task deep neural network, denoted as M-YOLACT, was designed to identify defects, lining profiles, and hyperbola shapes simultaneously. We introduced a curve-fitting post-processing operation to calculate the dielectric constant automatically based on the hyperbola shapes and evaluated the defect depths and lining thicknesses. The method was validated by numerical simulations, sandbox, and field tests. The method effectively identified the shapes and classes of tunnel defects as well as the thickness profiles from GPR B-Scan images.

MV-GPRNet: Multi-View Subsurface Defect Detection Network for Airport Runway Inspection Based on GPR

The detection and restoration of subsurface defects are essential for ensuring the structural reliability of airport runways. Subsurface inspections can be performed with the aid of a robot equipped with a Ground Penetrating Radar (GPR). However, interpreting GPR data is extremely difficult, as GPR data usually contains severe clutter interference. In addition, many different types of subsurface defects present similar features in B-scan images, making them difficult to distinguish. Consequently, this makes later maintenance work harder as different subsurface defects require different restoration measures. Thus, to automate the inspection process and improve defect identification accuracy, a novel deep learning algorithm, MV-GPRNet, is proposed. Instead of traditionally using GPR B-scan images only, MV-GPRNet utilizes multi-view GPR data to robustly detect regions with defects despite significant interference. It originally fuses the 3D feature map in C-scan data and the 2D feature map in Top-scan data for defect classification and localization. With our runway inspection robot, a large number of real runway data sets from three international airports have been used to extensively test our method. Experimental results indicate that the proposed MV-GPRNet outperforms state-of-the-art (SOTA) approaches. In particular, MV-GPRNet achieves F1 measurements for voids, cracks, subsidences, and pipes at 91%, 69%, 90%, and 100%, respectively.