Generation of High-Precision Ground Penetrating Radar Images Using Improved Least Square Generative Adversarial Networks

Yue, Yunpeng; Liu, Hai; Meng, Xu; Li, Yinguang; Du, Yanliang

doi:10.3390/rs13224590

Cited by 30 publications

(15 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It will even cover up effective signals. 2D GPR is not intuitive for imaging underground targets and can interfere with data processors' analysis to interpret information about the subsurface problem [11]. And the 2D radar images of some underground structures (such as trenches and pipes) are very similar to the 2D radar images of underground cavities, which can easily cause the problem of uncertainty of multiple different interpretations of the same radar image.…”

Section: Introductionmentioning

confidence: 99%

3D ground penetrating radar cavity identification algorithm for urban roads using transfer learning

Li¹,

Yang²,

Qiao³

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

3D ground penetrating radar (GPR) is the main method for the detection of underground cavities in urban roads. The number of road cavity samples detected by 3D radar is small, whereas the intelligent identification model require a large number of learning samples for model training, resulting in inadequate model training. This causes the model to be less accurate in identifying cavities, resulting in many misses and misjudgments. Given the above problems, combined with the detection characteristics of the vertical, the horizontal, and the crossed slices obtained in one detection process of 3D GPR, a 3D GPR cavity intelligent recognition model based on model-based transfer learning is proposed. Firstly, a large amount of 3D GPR data of urban road models with cavities are obtained through forwarding simulation. And the intelligent recognition model was pre-trained on the cavity detection data on three types of slices respectively. Then, through model-based transfer learning, a small amount of real underground cavity data is used to speed up the convergence speed of model training and optimize the structural parameters. It breaks through the limitation of the insufficient number of cavity samples for 3D radar detection on the intelligent learning model training, reduces algorithm training costs, and improves identification accuracy.

show abstract

Section: Introductionmentioning

confidence: 99%

3D ground penetrating radar cavity identification algorithm for urban roads using transfer learning

Li¹,

Yang²,

Qiao³

et al. 2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…To address the issue of insufficient real GPR images for network training, generative adversarial networks (GANs) 16 provide a new idea for generating GPR images. Yue et al 17 proposed an improved least squares generative adversarial network (LSGAN) model for generation of high-precision GPR images to address the scarcity of labeled GPR data. They verified that the inclusion of LSGAN-generated images in the training GPR dataset could increase the target diversity and improve the detection precision by 10% compared with the model trained on the dataset containing 500 field GPR images.…”

Section: Introductionmentioning

confidence: 99%

“…19 As a consequence, the generated images could be exactly the same as the real ones, and their diversity was decreased. 17 Another common method is to use the finite-difference time-domain (FDTD) 20 to generate synthesized GPR images. The simulated data generated by FDTD simulations have underground model labels, and the simulation can generate simulated data in a variety of scenarios.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised deep learning-based ground penetrating radar image translation for internal defect recognition of underground engineering structures

Wang

Lei

Jing

et al. 2023

Structural Health Monitoring

View full text Add to dashboard Cite

Anomaly detection of internal defects in underground engineering structures is critical. This paper proposes an unsupervised deep learning image-to-image translation method tailored for ground penetrating radar (GPR) images. The proposed model can translate real-world GPR images to simulated ones. In this manner, labeling real GPR images is not necessary, and only the target detection model trained on simulated GPR images is required to directly identify defects in real GPR images. The unsupervised deep learning network introduces geometry-consistency constraints into the CycleGAN, which largely prevents the problem of semantic distortion in translation. Validation of the proposed method was performed using GPR data collected in various scenarios using GPR of different center frequencies and manufacturers. Moreover, to verify its adaptability and feasibility for defect recognition, commonly used deep learning-based defect recognition methods, which were trained only on simulated GPR images, were used to detect the translated GPR images. The findings indicate that the type and location of internal defects in translated GPR images can be accurately identified using the proposed method.

show abstract

“…Compared with other traditional non-destructive testing (NDT) methods, ground penetrating radar (GPR) has been widely used in the detection of hidden defects due to its advantage of high-efficiency, high-resolution, and portability [4]. However, manual interpretation of GPR data is both time and labor consuming, especially for large amounts of field GPR data [5].…”

Section: Introductionmentioning

confidence: 99%

Automatic recognition of hidden defects behind railway tunnel lining using ground penetrating radar and deep learning

Yue

Liu

Lai

et al. 2022

International Conference on Smart Transportation and City Engineering (STCE 2022)

View full text Add to dashboard Cite

As a recognized non-destructive testing method, ground penetrating radar (GPR) has been widely applied for hidden defect detection of railway tunnels. However, the manual interpretation of GPR data is time-consuming and has a high demand of the operator’s experience. This paper presents an automatic recognition algorithm, which is based on a faster region-based convolutional neural network (Faster R-CNN), to recognize the hidden defects behind the railway tunnel lining, including non-compactness, voids with air- or water-filled, and cracks. To improve the recognition ability of the proposed algorithm, transfer learning is used by pre-training the network based on two groups of GPR dataset. Field experiment results preliminarily show that the proposed recognition algorithm can accurately identify and classify the hidden defects in the field GPR data.

show abstract

Generation of High-Precision Ground Penetrating Radar Images Using Improved Least Square Generative Adversarial Networks

Cited by 30 publications

References 36 publications

3D ground penetrating radar cavity identification algorithm for urban roads using transfer learning

3D ground penetrating radar cavity identification algorithm for urban roads using transfer learning

Unsupervised deep learning-based ground penetrating radar image translation for internal defect recognition of underground engineering structures

Automatic recognition of hidden defects behind railway tunnel lining using ground penetrating radar and deep learning

Contact Info

Product

Resources

About