A Strong Noise Reduction Network for Seismic Records

Noise reduction is a critical step in seismic data processing. A novel strong noise reduction network is proposed in this study. The network enhances the U-Net architecture with an improved inception module and coordinate attention (CA) mechanism, suppressing noise and enhancing signal clarity. Thes...

Full description

Saved in:
Bibliographic Details
Main Authors: Tong Shen, Xuan Jiang, Wenzheng Rong, Lei Xu, Xianguo Tuo, Guili Peng
Format: Article
Language:English
Published: MDPI AG 2024-11-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/14/22/10262
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850149896031567872
author Tong Shen
Xuan Jiang
Wenzheng Rong
Lei Xu
Xianguo Tuo
Guili Peng
author_facet Tong Shen
Xuan Jiang
Wenzheng Rong
Lei Xu
Xianguo Tuo
Guili Peng
author_sort Tong Shen
collection DOAJ
description Noise reduction is a critical step in seismic data processing. A novel strong noise reduction network is proposed in this study. The network enhances the U-Net architecture with an improved inception module and coordinate attention (CA) mechanism, suppressing noise and enhancing signal clarity. These enhancements improve the network’s capability to distinguish between signal and noise in the time–frequency domain. We trained and tested our model on the STEAD dataset, which eliminated noise across various frequency bands, improved the signal-to-noise ratio (SNR) of seismic records, and reduced the waveform distortion significantly. Comparative analyses against U-Net, DeepDenoiser, and DnRDB models, using signals with SNRs ranging from −14 dB to 0 dB, demonstrated our model’s superior performance. At the same time, we demonstrated that the Inception Conv Block has a significant impact on the denoising ability of the network. Furthermore, validation using the “Di Ting” dataset and real noisy signals confirmed the model’s generalizability. These results show that the proposed model significantly outperforms the comparative methods in terms of the SNR, correlation coefficient (r), and root mean square error (RMSE), delivering higher-quality seismograms. The enhanced phase-picking accuracy underscores the potential of our approach to advance in geophysics applications.
format Article
id doaj-art-efcf1d2e378b43f39f5e3718dd8eec4a
institution OA Journals
issn 2076-3417
language English
publishDate 2024-11-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-efcf1d2e378b43f39f5e3718dd8eec4a2025-08-20T02:26:45ZengMDPI AGApplied Sciences2076-34172024-11-0114221026210.3390/app142210262A Strong Noise Reduction Network for Seismic RecordsTong Shen0Xuan Jiang1Wenzheng Rong2Lei Xu3Xianguo Tuo4Guili Peng5School of Information Engineering, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Information Engineering, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Automation and Information Engineering, Sichuan University of Science and Engineering, Yibin 644000, ChinaSchool of Information Engineering, Southwest University of Science and Technology, Mianyang 621010, ChinaSchool of Automation and Information Engineering, Sichuan University of Science and Engineering, Yibin 644000, ChinaSchool of Control and Mechanical Engineering, Tianjin Chengjian University, Tianjin 300384, ChinaNoise reduction is a critical step in seismic data processing. A novel strong noise reduction network is proposed in this study. The network enhances the U-Net architecture with an improved inception module and coordinate attention (CA) mechanism, suppressing noise and enhancing signal clarity. These enhancements improve the network’s capability to distinguish between signal and noise in the time–frequency domain. We trained and tested our model on the STEAD dataset, which eliminated noise across various frequency bands, improved the signal-to-noise ratio (SNR) of seismic records, and reduced the waveform distortion significantly. Comparative analyses against U-Net, DeepDenoiser, and DnRDB models, using signals with SNRs ranging from −14 dB to 0 dB, demonstrated our model’s superior performance. At the same time, we demonstrated that the Inception Conv Block has a significant impact on the denoising ability of the network. Furthermore, validation using the “Di Ting” dataset and real noisy signals confirmed the model’s generalizability. These results show that the proposed model significantly outperforms the comparative methods in terms of the SNR, correlation coefficient (r), and root mean square error (RMSE), delivering higher-quality seismograms. The enhanced phase-picking accuracy underscores the potential of our approach to advance in geophysics applications.https://www.mdpi.com/2076-3417/14/22/10262denoiselow SNRdeep learningU-Netinceptionchannel attention
spellingShingle Tong Shen
Xuan Jiang
Wenzheng Rong
Lei Xu
Xianguo Tuo
Guili Peng
A Strong Noise Reduction Network for Seismic Records
Applied Sciences
denoise
low SNR
deep learning
U-Net
inception
channel attention
title A Strong Noise Reduction Network for Seismic Records
title_full A Strong Noise Reduction Network for Seismic Records
title_fullStr A Strong Noise Reduction Network for Seismic Records
title_full_unstemmed A Strong Noise Reduction Network for Seismic Records
title_short A Strong Noise Reduction Network for Seismic Records
title_sort strong noise reduction network for seismic records
topic denoise
low SNR
deep learning
U-Net
inception
channel attention
url https://www.mdpi.com/2076-3417/14/22/10262
work_keys_str_mv AT tongshen astrongnoisereductionnetworkforseismicrecords
AT xuanjiang astrongnoisereductionnetworkforseismicrecords
AT wenzhengrong astrongnoisereductionnetworkforseismicrecords
AT leixu astrongnoisereductionnetworkforseismicrecords
AT xianguotuo astrongnoisereductionnetworkforseismicrecords
AT guilipeng astrongnoisereductionnetworkforseismicrecords
AT tongshen strongnoisereductionnetworkforseismicrecords
AT xuanjiang strongnoisereductionnetworkforseismicrecords
AT wenzhengrong strongnoisereductionnetworkforseismicrecords
AT leixu strongnoisereductionnetworkforseismicrecords
AT xianguotuo strongnoisereductionnetworkforseismicrecords
AT guilipeng strongnoisereductionnetworkforseismicrecords