Chengyun Song, Shutao Guo, Chuanchao Xiong, Jiying Tuo
{"title":"Regularized deep learning for unsupervised random noise attenuation in poststack seismic data","authors":"Chengyun Song, Shutao Guo, Chuanchao Xiong, Jiying Tuo","doi":"10.1093/jge/gxad094","DOIUrl":null,"url":null,"abstract":"Deep learning methods achieve excellent noise reduction performances in seismic data processing compared with traditional methods. However, deep learning usually requires a large number of pairwise noisy-clean training data, which is an extremely challenging task. In this paper, an unsupervised approach without clean seismic data is proposed to suppress random noise. Seismic data is divided into odd and even traces, which serve as the input and output of the depth network. So that the proposed algorithm can be trained directly on the original data. What is more, the proposed method introduces two regularization terms to solve the over-smoothing problem caused by reconstruction of adjacent traces. The first term considers an ideal denoising network that does not cause oversmooth as a constraint, while the second term considers the structural information existing in seismic data. Experiments on synthetic post-stack data illustrate that the proposed method obtain the higher SNR than the comparison methods. In the application of field post-stack seismic data, the proposed method can effectively maintain the seismic amplitude and generate good spectral characteristics.","PeriodicalId":54820,"journal":{"name":"Journal of Geophysics and Engineering","volume":null,"pages":null},"PeriodicalIF":1.6000,"publicationDate":"2023-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Geophysics and Engineering","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1093/jge/gxad094","RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"GEOCHEMISTRY & GEOPHYSICS","Score":null,"Total":0}
引用次数: 0
Abstract
Deep learning methods achieve excellent noise reduction performances in seismic data processing compared with traditional methods. However, deep learning usually requires a large number of pairwise noisy-clean training data, which is an extremely challenging task. In this paper, an unsupervised approach without clean seismic data is proposed to suppress random noise. Seismic data is divided into odd and even traces, which serve as the input and output of the depth network. So that the proposed algorithm can be trained directly on the original data. What is more, the proposed method introduces two regularization terms to solve the over-smoothing problem caused by reconstruction of adjacent traces. The first term considers an ideal denoising network that does not cause oversmooth as a constraint, while the second term considers the structural information existing in seismic data. Experiments on synthetic post-stack data illustrate that the proposed method obtain the higher SNR than the comparison methods. In the application of field post-stack seismic data, the proposed method can effectively maintain the seismic amplitude and generate good spectral characteristics.
期刊介绍:
Journal of Geophysics and Engineering aims to promote research and developments in geophysics and related areas of engineering. It has a predominantly applied science and engineering focus, but solicits and accepts high-quality contributions in all earth-physics disciplines, including geodynamics, natural and controlled-source seismology, oil, gas and mineral exploration, petrophysics and reservoir geophysics. The journal covers those aspects of engineering that are closely related to geophysics, or on the targets and problems that geophysics addresses. Typically, this is engineering focused on the subsurface, particularly petroleum engineering, rock mechanics, geophysical software engineering, drilling technology, remote sensing, instrumentation and sensor design.