GPR Data Reconstruction Using Residual Feature Distillation Block U-Net
Due to the unevenness of ground surface, mismatch between trig interval and sampling speed, or other electromagnetic interferences, traces missing is a quite typical occurrence during the on-ground ground penetrating radar (GPR) testing. Effective reconstruction of GPR missing traces has been regard...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2023-01-01
|
Series: | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/10124341/ |
_version_ | 1797330411069636608 |
---|---|
author | Qianwei Dai Yue He Yi Lei Jianwei Lei Xiangyu Wang Bin Zhang |
author_facet | Qianwei Dai Yue He Yi Lei Jianwei Lei Xiangyu Wang Bin Zhang |
author_sort | Qianwei Dai |
collection | DOAJ |
description | Due to the unevenness of ground surface, mismatch between trig interval and sampling speed, or other electromagnetic interferences, traces missing is a quite typical occurrence during the on-ground ground penetrating radar (GPR) testing. Effective reconstruction of GPR missing traces has been regarded a crucial link to improve both the signal-to-noise ratio of raw data and the resolution of GPR imaging. In this article, we propose a novel deep-learning framework based on the residual feature distillation block U-Net (RFDB-U-Net) to mitigate the transmission loss problem of the conventional U-Net. To be specific, by employing the information distillation network based on the multiple feature extraction connections, RFDB is capable of utilizing the adequate residual information of each layer for feature learning. Moreover, a skip connection is additional patched on the residual units to properly compensate the missing features in the convolution process. In particular, the merging of lightweight U-Net ensures the lightness of RFDB. The outperformance of the proposed framework is verified in detail through the reconstruction accuracy and evaluation metrics in the test of synthetic data, laboratorial data, and in-site field data. |
first_indexed | 2024-03-08T07:19:27Z |
format | Article |
id | doaj.art-8736f993f5c443239293569bbc61f02b |
institution | Directory Open Access Journal |
issn | 2151-1535 |
language | English |
last_indexed | 2024-03-08T07:19:27Z |
publishDate | 2023-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing |
spelling | doaj.art-8736f993f5c443239293569bbc61f02b2024-02-03T00:01:30ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing2151-15352023-01-01166958696810.1109/JSTARS.2023.327616110124341GPR Data Reconstruction Using Residual Feature Distillation Block U-NetQianwei Dai0Yue He1https://orcid.org/0000-0001-5994-4928Yi Lei2https://orcid.org/0000-0001-6742-3823Jianwei Lei3https://orcid.org/0000-0002-1066-2302Xiangyu Wang4https://orcid.org/0000-0003-4078-3620Bin Zhang5https://orcid.org/0000-0002-2127-9560Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring, Ministry of Education, Central South University, Changsha, ChinaKey Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring, Ministry of Education, Central South University, Changsha, ChinaSchool of Civil Engineering, Central South University, Changsha, ChinaYellow River Laboratory, Zhengzhou University, Zhengzhou, ChinaKey Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring, Ministry of Education, Central South University, Changsha, ChinaKey Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring, Ministry of Education, Central South University, Changsha, ChinaDue to the unevenness of ground surface, mismatch between trig interval and sampling speed, or other electromagnetic interferences, traces missing is a quite typical occurrence during the on-ground ground penetrating radar (GPR) testing. Effective reconstruction of GPR missing traces has been regarded a crucial link to improve both the signal-to-noise ratio of raw data and the resolution of GPR imaging. In this article, we propose a novel deep-learning framework based on the residual feature distillation block U-Net (RFDB-U-Net) to mitigate the transmission loss problem of the conventional U-Net. To be specific, by employing the information distillation network based on the multiple feature extraction connections, RFDB is capable of utilizing the adequate residual information of each layer for feature learning. Moreover, a skip connection is additional patched on the residual units to properly compensate the missing features in the convolution process. In particular, the merging of lightweight U-Net ensures the lightness of RFDB. The outperformance of the proposed framework is verified in detail through the reconstruction accuracy and evaluation metrics in the test of synthetic data, laboratorial data, and in-site field data.https://ieeexplore.ieee.org/document/10124341/Deep learningground penetrating radar (GPR)missing tracesreconstructionresidual feature distillation block U-Net (RFDB-U-net) |
spellingShingle | Qianwei Dai Yue He Yi Lei Jianwei Lei Xiangyu Wang Bin Zhang GPR Data Reconstruction Using Residual Feature Distillation Block U-Net IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing Deep learning ground penetrating radar (GPR) missing traces reconstruction residual feature distillation block U-Net (RFDB-U-net) |
title | GPR Data Reconstruction Using Residual Feature Distillation Block U-Net |
title_full | GPR Data Reconstruction Using Residual Feature Distillation Block U-Net |
title_fullStr | GPR Data Reconstruction Using Residual Feature Distillation Block U-Net |
title_full_unstemmed | GPR Data Reconstruction Using Residual Feature Distillation Block U-Net |
title_short | GPR Data Reconstruction Using Residual Feature Distillation Block U-Net |
title_sort | gpr data reconstruction using residual feature distillation block u net |
topic | Deep learning ground penetrating radar (GPR) missing traces reconstruction residual feature distillation block U-Net (RFDB-U-net) |
url | https://ieeexplore.ieee.org/document/10124341/ |
work_keys_str_mv | AT qianweidai gprdatareconstructionusingresidualfeaturedistillationblockunet AT yuehe gprdatareconstructionusingresidualfeaturedistillationblockunet AT yilei gprdatareconstructionusingresidualfeaturedistillationblockunet AT jianweilei gprdatareconstructionusingresidualfeaturedistillationblockunet AT xiangyuwang gprdatareconstructionusingresidualfeaturedistillationblockunet AT binzhang gprdatareconstructionusingresidualfeaturedistillationblockunet |