Mohamed Lokman Mohd Yusof
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Mohamed Lokman Mohd Yusof has a total of 40 co-authors. Here are the top co-authors by number of shared publications:
- Abdulkadir C. Yucel
Nanyang Technological University
22 shared publications - Yee Hui Lee
School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore
22 shared publications - Hai-Han Sun
Department of Electrical and Computer Engineering, University of WisconsinMadison, Madison, WI, USA
10 shared publications - Wenhao Luo
College of Surveying and Geo-Informatics, Tongji University, Shanghai, China
10 shared publications - Qiqi Dai
School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore
9 shared publications - Lai Fern Ow
National Parks Board, Singapore
7 shared publications
Latest Publications
- 9 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-18
A Migration-Assisted Deep Learning Scheme for Imaging Defects Inside Cylindrical Structures via GPR: A Case Study for Tree Trunks
Ground-penetrating radar (GPR) has emerged as a prominent tool for imaging internal defects in cylindrical structures, such as columns, utility poles, and tree trunks. However, accurately reconstructing both the shape and permittivity of the defects inside cylindrical structures remains challenging due to complex wave scattering phenomena and the limited accuracy of the existing signal processing and deep learning techniques. To address these issues, this study proposes a migration-assisted deep learning scheme for reconstructing the shape and permittivity of defects within cylindrical structures. The proposed scheme involves three stages of GPR data processing. First, a dual-permittivity estimation network (DPE-Net) extracts the permittivity values of the defect and the cylindrical structure, the latter of which is estimated with the help of a novel structural similarity index measure (SSIM)-based autofocusing technique (AFT). Second, a modified Kirchhoff migration incorporating the extracted permittivity of the cylindrical structure maps the signals reflected from the defect to the imaging domain. Third, a shape reconstruction network (SR-Net) processes the migrated image to recover the precise shape of the defect. The image of the interior defect is finally obtained by combining the reconstructed shape and extracted permittivity of the defect. The proposed scheme is validated using both synthetic and experimental data from a laboratory trunk model and real tree trunk samples. Comparative results show superior performance over existing deep learning methods, while generalization tests on live trees confirm its feasibility for in-field deployment. The underlying principle can further be applied to other circumferential GPR imaging scenarios. The code and database are available at https://github.com/jwqian54/Migration-Assisted-DL
- 9 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-18
A Migration-Assisted Deep Learning Scheme for Imaging Defects Inside Cylindrical Structures via GPR: A Case Study for Tree Trunks
Ground-penetrating radar (GPR) has emerged as a prominent tool for imaging internal defects in cylindrical structures, such as columns, utility poles, and tree trunks. However, accurately reconstructing both the shape and permittivity of the defects inside cylindrical structures remains challenging due to complex wave scattering phenomena and the limited accuracy of the existing signal processing and deep learning techniques. To address these issues, this study proposes a migration-assisted deep learning scheme for reconstructing the shape and permittivity of defects within cylindrical structures. The proposed scheme involves three stages of GPR data processing. First, a dual-permittivity estimation network (DPE-Net) extracts the permittivity values of the defect and the cylindrical structure, the latter of which is estimated with the help of a novel structural similarity index measure (SSIM)-based autofocusing technique (AFT). Second, a modified Kirchhoff migration incorporating the extracted permittivity of the cylindrical structure maps the signals reflected from the defect to the imaging domain. Third, a shape reconstruction network (SR-Net) processes the migrated image to recover the precise shape of the defect. The image of the interior defect is finally obtained by combining the reconstructed shape and extracted permittivity of the defect. The proposed scheme is validated using both synthetic and experimental data from a laboratory trunk model and real tree trunk samples. Comparative results show superior performance over existing deep learning methods, while generalization tests on live trees confirm its feasibility for in-field deployment. The underlying principle can further be applied to other circumferential GPR imaging scenarios. The code and database are available at https://github.com/jwqian54/Migration-Assisted-DL
- Open Access
A Migration-Assisted Deep Learning Scheme for Imaging Defects Inside Cylindrical Structures via GPR: A Case Study for Tree Trunks
Ground-penetrating radar (GPR) has emerged as a prominent tool for imaging internal defects in cylindrical structures, such as columns, utility poles, and tree trunks. However, accurately reconstructing both the shape and permittivity of the defects inside cylindrical structures remains challenging due to complex wave scattering phenomena and the limited accuracy of the existing signal processing and deep learning techniques. To address these issues, this study proposes a migration-assisted deep learning scheme for reconstructing the shape and permittivity of defects within cylindrical structures. The proposed scheme involves three stages of GPR data processing. First, a dual-permittivity estimation network extracts the permittivity values of the defect and the cylindrical structure, the latter of which is estimated with the help of a novel structural similarity index measure-based autofocusing technique. Second, a modified Kirchhoff migration incorporating the extracted permittivity of the cylindrical structure maps the signals reflected from the defect to the imaging domain. Third, a shape reconstruction network processes the migrated image to recover the precise shape of the defect. The image of the interior defect is finally obtained by combining the reconstructed shape and extracted permittivity of the defect. The proposed scheme is validated using both synthetic and experimental data from a laboratory trunk model and real tree trunk samples. Comparative results show superior performance over existing deep learning methods, while generalization tests on live trees confirm its feasibility for in-field deployment. The underlying principle can further be applied to other circumferential GPR imaging scenarios. The code and database are available at: https://github.com/jwqian54/Migration-Assisted-DL.
Open AccessA Migration-Assisted Deep Learning Scheme for Imaging Defects Inside Cylindrical Structures via GPR: A Case Study for Tree Trunks
Ground-penetrating radar (GPR) has emerged as a prominent tool for imaging internal defects in cylindrical structures, such as columns, utility poles, and tree trunks. However, accurately reconstructing both the shape and permittivity of the defects inside cylindrical structures remains challenging due to complex wave scattering phenomena and the limited accuracy of the existing signal processing and deep learning techniques. To address these issues, this study proposes a migration-assisted deep learning scheme for reconstructing the shape and permittivity of defects within cylindrical structures. The proposed scheme involves three stages of GPR data processing. First, a dual-permittivity estimation network extracts the permittivity values of the defect and the cylindrical structure, the latter of which is estimated with the help of a novel structural similarity index measure-based autofocusing technique. Second, a modified Kirchhoff migration incorporating the extracted permittivity of the cylindrical structure maps the signals reflected from the defect to the imaging domain. Third, a shape reconstruction network processes the migrated image to recover the precise shape of the defect. The image of the interior defect is finally obtained by combining the reconstructed shape and extracted permittivity of the defect. The proposed scheme is validated using both synthetic and experimental data from a laboratory trunk model and real tree trunk samples. Comparative results show superior performance over existing deep learning methods, while generalization tests on live trees confirm its feasibility for in-field deployment. The underlying principle can further be applied to other circumferential GPR imaging scenarios. The code and database are available at: https://github.com/jwqian54/Migration-Assisted-DL.
- 2 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-18
Automatic Dual-Polarized Ground Penetrating Radar for Enhanced 3-D Tree Roots System Architecture Reconstruction
Tree root systems are crucial for providing structural support and stability to trees. However, in urban environments, they can pose challenges due to potential conflicts with the foundations of roads and infrastructure, leading to significant damage. Therefore, there is a pressing need to investigate the subsurface tree root system architecture (RSA). Ground Penetrating Radar (GPR) has emerged as a powerful tool for this purpose, offering high-resolution and non-destructive testing capabilities. One of the primary challenges in enhancing GPR’s ability to detect roots lies in accurately reconstructing the three-dimensional (3D) structure of complex RSAs. This challenge is exacerbated by subsurface heterogeneity and intricate interlacement of root branches, which can result in erroneous stacking of two-dimensional (2D) root points during 3D reconstruction. This study introduces a novel approach employing our developed wheel-based dual-polarized GPR system capable of capturing four polarimetric scattering parameters at each scan point through automated zigzag movements. A dedicated radar signal processing framework analyzes these dual-polarized signals to extract essential root parameters. These parameters are then utilized in an optimized slice relation clustering (OSRC) algorithm, specifically designed for improving the reconstruction of complex RSA. The efficacy of integrating root parameters derived from dual-polarized GPR signals into the OSRC algorithm is initially evaluated through simulations to assess its capability in RSA reconstruction. Subsequently, the GPR system and processing methodology are validated under real-world conditions using natural Angsana tree root systems. The findings demonstrate a promising methodology for enhancing the accurate reconstruction of intricate 3D tree RSA structures.
- 2 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-18
Automatic Dual-Polarized Ground Penetrating Radar for Enhanced 3-D Tree Roots System Architecture Reconstruction
Tree root systems are crucial for providing structural support and stability to trees. However, in urban environments, they can pose challenges due to potential conflicts with the foundations of roads and infrastructure, leading to significant damage. Therefore, there is a pressing need to investigate the subsurface tree root system architecture (RSA). Ground Penetrating Radar (GPR) has emerged as a powerful tool for this purpose, offering high-resolution and non-destructive testing capabilities. One of the primary challenges in enhancing GPR’s ability to detect roots lies in accurately reconstructing the three-dimensional (3D) structure of complex RSAs. This challenge is exacerbated by subsurface heterogeneity and intricate interlacement of root branches, which can result in erroneous stacking of two-dimensional (2D) root points during 3D reconstruction. This study introduces a novel approach employing our developed wheel-based dual-polarized GPR system capable of capturing four polarimetric scattering parameters at each scan point through automated zigzag movements. A dedicated radar signal processing framework analyzes these dual-polarized signals to extract essential root parameters. These parameters are then utilized in an optimized slice relation clustering (OSRC) algorithm, specifically designed for improving the reconstruction of complex RSA. The efficacy of integrating root parameters derived from dual-polarized GPR signals into the OSRC algorithm is initially evaluated through simulations to assess its capability in RSA reconstruction. Subsequently, the GPR system and processing methodology are validated under real-world conditions using natural Angsana tree root systems. The findings demonstrate a promising methodology for enhancing the accurate reconstruction of intricate 3D tree RSA structures.
- Open Access
- 25 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 17 (19391404) , 19668-19681
Learning From Clutter: An Unsupervised Learning-Based Clutter Removal Scheme for GPR B-Scans
Ground-penetrating radar (GPR) data are often contaminated by hardware and environmental clutter, which significantly affects the accuracy and reliability of target response identification. Existing supervised deep learning techniques for removing clutter in GPR data require generating a large set of clutter-free B-scans as labels for training, which are computationally expensive in simulation and unfeasible in real-world experiments. To tackle this issue, we propose a two-stage unsupervised learning-based clutter removal scheme, called ULCR-Net, to obtain clutter-free GPR B-scans. In the first stage of the proposed scheme, a diffusion model tailored for GPR data augmentation is employed to generate a diverse set of raw B-scans from the input random noise. With the augmented dataset, the second stage of the proposed scheme uses a contrastive learning-based generative adversarial network to learn and estimate clutter patterns in the raw B-scan. The clutter-free B-scan is then obtained by subtracting the clutter pattern from the raw B-scan. The training of the two-stage network only requires a small set of raw B-scans and clutter-only B-scans that are readily available in real-world applications. Extensive experiments have been conducted to validate the effectiveness of the proposed method. Results on simulation and measurement data demonstrate that the proposed method has superior clutter removal accuracy and generalizability and outperforms existing algebraic techniques and supervised learning-based methods with limited training data by a large margin. With its high clutter suppression capability and low training data requirements, the proposed method is well-suited to remove clutter and restore target responses in real-world GPR applications.
- 25 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
- Vol. 17 (19391404) , 19668-19681
Open AccessLearning From Clutter: An Unsupervised Learning-Based Clutter Removal Scheme for GPR B-Scans
Ground-penetrating radar (GPR) data are often contaminated by hardware and environmental clutter, which significantly affects the accuracy and reliability of target response identification. Existing supervised deep learning techniques for removing clutter in GPR data require generating a large set of clutter-free B-scans as labels for training, which are computationally expensive in simulation and unfeasible in real-world experiments. To tackle this issue, we propose a two-stage unsupervised learning-based clutter removal scheme, called ULCR-Net, to obtain clutter-free GPR B-scans. In the first stage of the proposed scheme, a diffusion model tailored for GPR data augmentation is employed to generate a diverse set of raw B-scans from the input random noise. With the augmented dataset, the second stage of the proposed scheme uses a contrastive learning-based generative adversarial network to learn and estimate clutter patterns in the raw B-scan. The clutter-free B-scan is then obtained by subtracting the clutter pattern from the raw B-scan. The training of the two-stage network only requires a small set of raw B-scans and clutter-only B-scans that are readily available in real-world applications. Extensive experiments have been conducted to validate the effectiveness of the proposed method. Results on simulation and measurement data demonstrate that the proposed method has superior clutter removal accuracy and generalizability and outperforms existing algebraic techniques and supervised learning-based methods with limited training data by a large margin. With its high clutter suppression capability and low training data requirements, the proposed method is well-suited to remove clutter and restore target responses in real-world GPR applications.
- Open Access
A Compact Dual-Polarized Vivaldi Antenna with High Gain for Tree Radar Applications
A dual-polarized compact Vivaldi antenna with high gain performance is proposed for tree radar applications. The proposed design introduces an array configuration consisting of two pairs of two Vivaldi elements to optimize the operating bandwidth and gain while providing dual-polarization capability. To enhance the gain of the proposed antenna over a certain frequency range of interest, directors and edge slots are incorporated into each Vivaldi element. To further enhance the overall antenna gain, a metal back reflector is used. The measurement results of the fabricated antenna show that the proposed antenna achieves a high gain of 5.5 to 14.8 dBi over broadband from 0.5 GHz to 3 GHz. Moreover, it achieves cross-polarization discrimination larger than 20 dB, ensuring high polarization purity. The fabricated antenna is used to detect and image the defects inside tree trunks. The results show that the proposed antenna yields a better-migrated image with a clear defect region compared to that obtained by a commercial Horn antenna.
Open AccessA Compact Dual-Polarized Vivaldi Antenna with High Gain for Tree Radar Applications
A dual-polarized compact Vivaldi antenna with high gain performance is proposed for tree radar applications. The proposed design introduces an array configuration consisting of two pairs of two Vivaldi elements to optimize the operating bandwidth and gain while providing dual-polarization capability. To enhance the gain of the proposed antenna over a certain frequency range of interest, directors and edge slots are incorporated into each Vivaldi element. To further enhance the overall antenna gain, a metal back reflector is used. The measurement results of the fabricated antenna show that the proposed antenna achieves a high gain of 5.5 to 14.8 dBi over broadband from 0.5 GHz to 3 GHz. Moreover, it achieves cross-polarization discrimination larger than 20 dB, ensuring high polarization purity. The fabricated antenna is used to detect and image the defects inside tree trunks. The results show that the proposed antenna yields a better-migrated image with a clear defect region compared to that obtained by a commercial Horn antenna.
- 11 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-15
A Deep Learning-Augmented Stand-Off Radar Scheme for Rapidly Detecting Tree Defects
Tree defect detection is crucial for the structural health screening of trees. Existing nondestructive testing (NDT) techniques for tree defect detection require time-consuming and labor-intensive measurement campaigns. This discourages their application for the routine structural health screening of whole populations of managed urban trees. To address this issue, this study proposes a deep-learning augmented stand-off radar scheme for contactless scanning of tree trunks and rapid detection of tree defects. In this scheme, the antenna is moved along a straight trajectory at a distance from the tree trunk to obtain the trunk’s B-scan. The obtained raw B-scan is then processed by a signal-processing framework specifically developed for revealing the scattering signatures of defects in B-scan, which achieves a 30 dB and 22 dB increase in the signal-to-clutter and noise ratio of the measurement data of tree trunk samples and living trees, respectively. Finally, the processed B-scan is input into a multilevel feature fusion neural network particularly designed for extracting the signature of the defect in the processed B-scan in real time. The developed scheme’s applications to the detection of defects in real fresh-cut tree trunks show that the stand-off radar scheme can detect tree defects with 96% accuracy. This stand-off radar scheme is the first contactless NDT technique for tree defect detection while operated on a straight trajectory and potentially can be integrated into the routine tree inspection workflow which is part of urban tree management.
- 11 June 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 62 (01962892) , 1-15
A Deep Learning-Augmented Stand-Off Radar Scheme for Rapidly Detecting Tree Defects
Tree defect detection is crucial for the structural health screening of trees. Existing nondestructive testing (NDT) techniques for tree defect detection require time-consuming and labor-intensive measurement campaigns. This discourages their application for the routine structural health screening of whole populations of managed urban trees. To address this issue, this study proposes a deep-learning augmented stand-off radar scheme for contactless scanning of tree trunks and rapid detection of tree defects. In this scheme, the antenna is moved along a straight trajectory at a distance from the tree trunk to obtain the trunk’s B-scan. The obtained raw B-scan is then processed by a signal-processing framework specifically developed for revealing the scattering signatures of defects in B-scan, which achieves a 30 dB and 22 dB increase in the signal-to-clutter and noise ratio of the measurement data of tree trunk samples and living trees, respectively. Finally, the processed B-scan is input into a multilevel feature fusion neural network particularly designed for extracting the signature of the defect in the processed B-scan in real time. The developed scheme’s applications to the detection of defects in real fresh-cut tree trunks show that the stand-off radar scheme can detect tree defects with 96% accuracy. This stand-off radar scheme is the first contactless NDT technique for tree defect detection while operated on a straight trajectory and potentially can be integrated into the routine tree inspection workflow which is part of urban tree management.
- 5 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-10
A Depth-Adaptive Filtering Method for Effective GPR Tree Roots Detection in Tropical Area
This study presents a technique for processing Step-frequency continuous wave (SFCW) ground penetrating radar (GPR) data to detect tree roots. SFCW GPR is portable and enables precise control of energy levels, balancing depth and resolution trade-offs. However, the high-frequency components of the transmission band suffers from poor penetrating capability and generates noise that interferes with root detection. The proposed time-frequency filtering technique uses a short-time Fourier transform (STFT) to track changes in frequency spectrum density over time. To obtain the filter window, a weighted linear regression (WLR) method is used. By adopting a conversion method that is a variant of the chirp Z-Transform (CZT), the time-frequency window filters out frequency samples that are not of interest when doing the frequency-to-time domain data conversion. The proposed depth-adaptive filter window can self-adjust to different scenarios, making it independent of soil information and effectively determines subsurface tree roots. The technique is successfully validated using SFCW GPR data from actual sites in a tropical area with different soil moisture levels, and the two-dimensional (2D) radar map of subsurface root systems is highly improved compared to existing methods.
- 5 June 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 72 (00189456) , 1-10
A Depth-Adaptive Filtering Method for Effective GPR Tree Roots Detection in Tropical Area
This study presents a technique for processing Step-frequency continuous wave (SFCW) ground penetrating radar (GPR) data to detect tree roots. SFCW GPR is portable and enables precise control of energy levels, balancing depth and resolution trade-offs. However, the high-frequency components of the transmission band suffers from poor penetrating capability and generates noise that interferes with root detection. The proposed time-frequency filtering technique uses a short-time Fourier transform (STFT) to track changes in frequency spectrum density over time. To obtain the filter window, a weighted linear regression (WLR) method is used. By adopting a conversion method that is a variant of the chirp Z-Transform (CZT), the time-frequency window filters out frequency samples that are not of interest when doing the frequency-to-time domain data conversion. The proposed depth-adaptive filter window can self-adjust to different scenarios, making it independent of soil information and effectively determines subsurface tree roots. The technique is successfully validated using SFCW GPR data from actual sites in a tropical area with different soil moisture levels, and the two-dimensional (2D) radar map of subsurface root systems is highly improved compared to existing methods.
- Open Access
A Depth-Adaptive Filtering Method for Effective GPR Tree Roots Detection in Tropical Area
This study presents a technique for processing Stepfrequency continuous wave (SFCW) ground penetrating radar (GPR) data to detect tree roots. SFCW GPR is portable and enables precise control of energy levels, balancing depth and resolution trade-offs. However, the high-frequency components of the transmission band suffers from poor penetrating capability and generates noise that interferes with root detection. The proposed time-frequency filtering technique uses a short-time Fourier transform (STFT) to track changes in frequency spectrum density over time. To obtain the filter window, a weighted linear regression (WLR) method is used. By adopting a conversion method that is a variant of the chirp Z-Transform (CZT), the timefrequency window filters out frequency samples that are not of interest when doing the frequency-to-time domain data conversion. The proposed depth-adaptive filter window can selfadjust to different scenarios, making it independent of soil information and effectively determines subsurface tree roots. The technique is successfully validated using SFCW GPR data from actual sites in a tropical area with different soil moisture levels, and the two-dimensional (2D) radar map of subsurface root systems is highly improved compared to existing methods.
Open AccessA Depth-Adaptive Filtering Method for Effective GPR Tree Roots Detection in Tropical Area
This study presents a technique for processing Stepfrequency continuous wave (SFCW) ground penetrating radar (GPR) data to detect tree roots. SFCW GPR is portable and enables precise control of energy levels, balancing depth and resolution trade-offs. However, the high-frequency components of the transmission band suffers from poor penetrating capability and generates noise that interferes with root detection. The proposed time-frequency filtering technique uses a short-time Fourier transform (STFT) to track changes in frequency spectrum density over time. To obtain the filter window, a weighted linear regression (WLR) method is used. By adopting a conversion method that is a variant of the chirp Z-Transform (CZT), the timefrequency window filters out frequency samples that are not of interest when doing the frequency-to-time domain data conversion. The proposed depth-adaptive filter window can selfadjust to different scenarios, making it independent of soil information and effectively determines subsurface tree roots. The technique is successfully validated using SFCW GPR data from actual sites in a tropical area with different soil moisture levels, and the two-dimensional (2D) radar map of subsurface root systems is highly improved compared to existing methods.
- Top Cited Papers
- 11 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-16
3DInvNet: A Deep Learning-Based 3D Ground-Penetrating Radar Data Inversion
The reconstruction of the 3D permittivity map from ground-penetrating radar (GPR) data is of great importance for mapping subsurface environments and inspecting underground structural integrity. Traditional iterative 3D reconstruction algorithms suffer from strong non-linearity, ill-posedness, and high computational cost. To tackle these issues, a 3D deep learning scheme, called 3DInvNet, is proposed to reconstruct 3D permittivity maps from GPR C-scans. The proposed scheme leverages a prior 3D convolutional neural network with a feature attention mechanism to suppress the noise in the C-scans due to subsurface heterogeneous soil environments. Then a 3D U-shaped encoder-decoder network with multi-scale feature aggregation modules is designed to establish the optimal inverse mapping from the denoised C-scans to 3D permittivity maps. Furthermore, a three-step separate learning strategy is employed to pre-train and fine-tune the networks. The proposed scheme is applied to numerical simulation as well as real measurement data. The quantitative and qualitative results show the networks’ capability, generalizability, and robustness in denoising GPR C-scans and reconstructing 3D permittivity maps of subsurface objects.
- 11 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-16
Top Cited Papers3DInvNet: A Deep Learning-Based 3D Ground-Penetrating Radar Data Inversion
The reconstruction of the 3D permittivity map from ground-penetrating radar (GPR) data is of great importance for mapping subsurface environments and inspecting underground structural integrity. Traditional iterative 3D reconstruction algorithms suffer from strong non-linearity, ill-posedness, and high computational cost. To tackle these issues, a 3D deep learning scheme, called 3DInvNet, is proposed to reconstruct 3D permittivity maps from GPR C-scans. The proposed scheme leverages a prior 3D convolutional neural network with a feature attention mechanism to suppress the noise in the C-scans due to subsurface heterogeneous soil environments. Then a 3D U-shaped encoder-decoder network with multi-scale feature aggregation modules is designed to establish the optimal inverse mapping from the denoised C-scans to 3D permittivity maps. Furthermore, a three-step separate learning strategy is employed to pre-train and fine-tune the networks. The proposed scheme is applied to numerical simulation as well as real measurement data. The quantitative and qualitative results show the networks’ capability, generalizability, and robustness in denoising GPR C-scans and reconstructing 3D permittivity maps of subsurface objects.
- 3 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-10
Slice-Relation-Clustering Framework via Horizontal Angle Information for 3-D Tree Roots Reconstruction
Tree root system 3D reconstruction and spatial distribution analysis are prevalent aspects of tree root investigation using ground penetrating radar (GPR). Precedent 3D reconstruction methods are found to be effective in mapping simple, smooth root structures. However, repetitive and dense B-scans are needed, otherwise, the retrieved roots’ spatial distribution and growth extension trend accuracy would deteriorate with the increase in the root systems’ complexity. To address these issues, this paper, for the first time, explores the possibility of integrating the horizontal angle information of the tree roots and a slice-relation-clustering (SRC) algorithm to reconstruct the complex tree root systems in a 3D manner. The proposed framework, which takes the roots’ horizontal angle as an analyzing condition instead of biological properties that are similar among neighboring branches used in existing methods, clusters pre-processed and focused 2D reflection patterns from the same single root together. The whole roots system is the combination of every single root cluster. Real measurement results show that our proposed method achieves a high efficiency in accurate root system reconstruction.
- 3 May 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 61 (01962892) , 1-10
Slice-Relation-Clustering Framework via Horizontal Angle Information for 3-D Tree Roots Reconstruction
Tree root system 3D reconstruction and spatial distribution analysis are prevalent aspects of tree root investigation using ground penetrating radar (GPR). Precedent 3D reconstruction methods are found to be effective in mapping simple, smooth root structures. However, repetitive and dense B-scans are needed, otherwise, the retrieved roots’ spatial distribution and growth extension trend accuracy would deteriorate with the increase in the root systems’ complexity. To address these issues, this paper, for the first time, explores the possibility of integrating the horizontal angle information of the tree roots and a slice-relation-clustering (SRC) algorithm to reconstruct the complex tree root systems in a 3D manner. The proposed framework, which takes the roots’ horizontal angle as an analyzing condition instead of biological properties that are similar among neighboring branches used in existing methods, clusters pre-processed and focused 2D reflection patterns from the same single root together. The whole roots system is the combination of every single root cluster. Real measurement results show that our proposed method achieves a high efficiency in accurate root system reconstruction.