Robust seismic data denoising based on deep learning
ZHANG Yan1, LI Xinyue1, WANG Bin1, LI Jie1, WANG Hongtao2, DONG Hongli3
1. School of Computer&Information Technology, Northeast Petroleum University, Daqing, Heilongjiang 163318, China; 2. Daqing Information Technology Research Center, Daqing, Heilongjiang 163318, China; 3. Artificial Intelligence Energy Research Institute, Northeast Petroleum University, Daqing, Heilongjiang 163318, China
Abstract:Noise in seismic data is complicated, and the traditional modeling methods based on prior knowledge cannot describe the noise distribution accurately. In the denoising methods based on deep learning, a multi-layer convolutional neural network is employed to automatically extract the deep features of seismic data, and its nonlinear approximation ability is used for adaptive learning, which yields a complex denoising model and thus brings a new idea for the denoising of seismic data. How-ever, poor generalization ability is found in the cur-rent denoising methods based on deep learning in the case of insufficient sample coverage, greatly reducing the denoising effect. Therefore, this paper proposes a robust deep learning algorithm for denoising. The model is composed of two sub-networks, which realize the estimation of noise distribution and noise suppression of noisy seismic data respectively. The sub-network for estimating noise distribution is a multi-layer convolutional neural network. The sub-network for denoising introduces a strategy of feature fusion, which comprehensively considers the global and local information of seismic data, and a residual learning strategy is utilized to extract noise features. L1 norm loss is taken as the loss function for the two sub-networks to enhance the generalization ability of the model. Experiments show that the method proposed in this paper has a higher generalization ability than similar algorithms. Data processing results indicate that it better preserves event features and has a higher signal-to-noise ratio.
张岩, 李新月, 王斌, 李杰, 王洪涛, 董宏丽. 基于深度学习的鲁棒地震数据去噪[J]. 石油地球物理勘探, 2022, 57(1): 12-25.
ZHANG Yan, LI Xinyue, WANG Bin, LI Jie, WANG Hongtao, DONG Hongli. Robust seismic data denoising based on deep learning. Oil Geophysical Prospecting, 2022, 57(1): 12-25.
张军华,吕宁,田连玉,等.地震资料去噪方法综合评述[J].石油地球物理勘探,2005,40(增刊1):121-127.ZHANG Junhua,LYU Ning,TIAN Lianyu,et al.A comprehensive review of seismic data denoising me-thods[J].Oil Geophysical Prospecting,2005,40(S1):121-127.
[2]
胡天跃.地震资料叠前去噪技术的现状与未来[J].地球物理学进展,2002,17(2):218-223.HU Tianyue.The current situation and future of seismic data prestack noise attenuation techniques[J].Progress in Geophysics,2002,17(2):218-223.
[3]
唐金良,曹辉,王立华,等.中值滤波在井间地震资料处理中的应用[J].石油物探,2005,44(1):47-50.TANG Jinliang,CAO Hui,WANG Lihua,et al.Median filter in seismic crosshole data processing[J].Geophysical Prospecting for Petroleum,2005, 44(1):47-50.
[4]
刘财,王典,刘洋,等.二维多级中值滤波技术在随机噪声消除中的应用初探[J].石油地球物理勘探,2005,40(2):163-167.LIU Cai,WANG Dian,LIU Yang,et al.Preliminary study of using 2D multi-level median filtering technique to eliminate random noises[J].Oil Geophysical Prospecting,2005,40(2):163-167.
[5]
王伟,高静怀,陈文超,等.基于结构自适应中值滤波器的随机噪声衰减方法[J].地球物理学报,2012,55(5):1732-1741.WANG Wei,GAO Jinghuai,CHEN Wenchao,et al.Random seismic noise suppression via structure-adaptive median filter[J].ChineseJournal of Geophysics,2012,55(5):1732-1741.
[6]
BUADES A,COLL B,MOREL J M.A non-local algorithm for image denoising[C].IEEE Computer Society Conference on Computer Vision and Pattern Recognition,2005,60-65.
[7]
BONAR D,SACCHI M.Denoising seismic data using the nonlocal means algorithm[J].Geophysics,2012,77(1):A5-A8.
[8]
ALSDORF D.Noise reduction in seismic data using Fourier correction coefficient filtering[J].Geophy-sics,1997,62(5):1617-1627.
[9]
TRAD D,ULRYCH T,SACCHI M.Latest view of sparse Radon transforms[J].Geophysics,2003,68(10):386-399.
[10]
CAO S,CHEN X.The second-generation wavelet transform and its application in denoising of seismic data[J].Applied Geophysics,2005,2(2):70-74.
[11]
高静怀,毛剑,满蔚仕,等.叠前地震资料噪声衰减的小波域方法研究[J].地球物理学报,2006,49(4):1155-1163.GAO Jinghuai,MAO Jian,MAN Weishi,et al.On the denoising method of prestack seismic data in wavelet domain[J].Chinese Journal of Geophysics,2006,49(4):1155-1163.
[12]
NEELAMANI R,BAUMSTEIN A I,GILARD D,et al.Coherent and random noise attenuation using the curvelet transform[J].The Leading Edge,2008,27(2):240-248.
[13]
ELAD M,AHARON M.Image denoising via sparse and redundant representations over learned dictionaries[J].IEEE Transactions on Image Processing,2006,15(12):3736-3745.
[14]
张岩,任伟建,唐国维.应用结构聚类字典学习压制地震数据随机噪声[J].石油地球物理勘探,2018,53(6):1119-1127.ZHANG Yan,REN Weijian,TANG Guowei.Random noise suppression on seismic data based on structured-clustering dictionary learning[J].Oil Geophysical Prospecting,2018,53(6):1119-1127.
[15]
ZHOU Y T, SHI C J, CHEN H M, et al.Spike-like blending noise attenuation using structural low-rank decomposition[J].IEEE Geoscience and Remote Sensing Letters, 2017,14(9):1633-1637.
[16]
ZU S H,ZHOU H,WU R S,et al.Hybrid-sparsity constrained dictionary learning for iterative deblen-ding of extremely noisy simultaneous-source data[J].IEEE Transactions on Geoscience and Remote Sen-sing,2019,57(4):2249-2262.
[17]
DABOV K,FOI A,KATKOVNIK V,et al.Image denoising by sparse 3-D transform-domain collaborative filtering[J].IEEE Transactions on Image Processing,2007,16(8):2080-2095.
[18]
ANVARI R,KAHOO A R,MOHAMMADI M,et al.Seismic random noise attenuation using sparse low-rank estimation of the signal in the time-frequency domain[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2019,12(5):1612-1618.
[19]
ZHOU Q B,GAO J H,WANG Z G,et al.Adaptive variable time fractional anisotropic diffusion filtering for seismic data noise attenuation[J].IEEE Transactions on Geoscience and Remote Sensing,2016,54(4):1905-1917.
[20]
ZHANG H,YANG H,LI H X,et al.Random noise attenuation of non-uniformly sampled 3D seismic data along two spatial coordinates using non-equispaced curvelet transform[J].Journal of Applied Geophy-sics,2018,151:221-233.
[21]
张玉玺,刘洋,张浩然,等.基于深度学习的多属性盐丘自动识别方法[J].石油地球物理勘探,2020,55(3):475-483.ZHANG Yuxi,LIU Yang,ZHANG Haoran,et al. Multi-attribute automatic interpretation of salt domes based on deep learning[J].Oil Geophysical Prospecting,2020,55(3):475-483.
[22]
王钰清,陆文凯,刘金林,等.基于数据增广和CNN的地震随机噪声压制[J].地球物理学报,2019,62(1):421-433.WANG Yuqing,LU Wenkai,LIU Jinlin,et al.Random seismic noise attenuation based on data augmentation and CNN[J].ChineseJournal of Geophysics,2019,62(1):421-433.
[23]
李海山,陈德武,吴杰,等.叠前随机噪声深度残差网络压制方法[J].石油地球物理勘探,2020,55(3):493-503.LI Haishan,CHEN Dewu,WU Jie,et al.Pre-stack random noise suppression with deep residual network[J].Oil Geophysical Prospecting,2020,55(3):493-503.
[24]
ZHANG M,LIU Y,CHEN Y K.Unsupervised seismic random noise attenuation based on deep convolutional neural network[J].IEEE Access,2019,7:179810-179822.
[25]
CHEN Y K,ZHANG M,BAI M,et al.Improving the signal-to-noise ratio of seismological datasets by unsupervised machine learning[J].Seismological Research Letters,2019,90(4):1552-1564.
[26]
罗仁泽,李阳阳.一种基于RUnet卷积神经网络的地震资料随机噪声压制方法[J].石油物探,2020,59(1):51-59.LUO Renze,LI Yangyang.Random seismic noise attenuation based on RUnet convolutional neural network[J].Geophysical Prospecting for Petroleum,2020,59(1):51-59.
[27]
ZHANG K,ZUO W M,CHEN Y J,et al.Beyond a Gaussian denoiser:Residual learning of deep CNN for image denoising[J].IEEE Transactions on Image Processing,2017, 26(7):3142-3155.
[28]
SHAN C H,GUO X R,QU J.Residual learning of deep convolutional neural networks for image denoi-sing[J].Intelligent & Fuzzy Systems,2019,37(2):2809-2818.
[29]
WU K L,ZHANG C H.Deep generative adversarial networks for the sparse signal denoising[C].International Conference on Pattern Recognition,2018,1127-1132.
[30]
韩卫雪,周亚同,池越.基于深度学习卷积神经网络的地震数据随机噪声去除[J].石油物探,2018,57(6):862-869.HAN Weixue,ZHOU Yatong,CHI Yue.Deep lear-ning convolutional neural networks for random noise attenuation in seismic data[J].Geophysical Prospecting for Petroleum,2018,57(6):862-869.
[31]
WANG F,CHEN S C.Residual learning of deep convolutional neural network for seismic random noise attenuation[J].IEEE Geoscience and Remote Sen-sing,2019,16(8):1314-1318.
[32]
YU S W,MA J W,WANG W L.Deep learning for denoising[J].Geophysics,2019,84(6):V333-V350.
[33]
ZHAO Y X,LI Y,DONG X T,et al.Low-frequency noise suppression method based on improved DnCNN in desert seismic data[J].IEEE Geoscience and Remote Sensing,2019,16(5):811-815.
[34]
DONG X T,LI Y,YANG B J.Desert low-frequency noise suppression by using adaptive DnCNNs based on the determination of high-order statistic[J].Geophysical Journal International,2019,219(2):1281-1299.
[35]
YANG L Q,CHEN W,LIU W,et al.Random noise attenuation based on residual convolutional neural network in seismic datasets[J].IEEE Access,2020,8:30271-30286.
[36]
GOODFELLOW I,POUGET-ABADIE J,MIRZA M,et al.Generative adversarial nets[C].Advances in Neural Information Processing Systems,2014,2672-2680.
[37]
RADFORD A,METZ L,CHINTALA S.Unsupervised representation learning with deep convolutional generative adversarial networks[C].International Conference on Learning Representations,2016,1-16.
[38]
俞若水,张勇,周创.基于深度卷积生成对抗网络的瑞雷波信号随机噪声去除[J].地球物理学进展,2020,35(6):2276-2283.YU Ruoshui,ZHANG Yong,ZHOU Chuang.Deep convolutional generative adversarial network for random noise attenuation in Rayleigh wave signal[J].Progress in Geophysics,2020,35(6):2276-2283.
[39]
MOHAN S, KADKHODAIE Z, SIMONCELLI E P,et al.Robust and interpretable blind image denoising via bias-free convolutional neural networks[C].International Conference on Learning Representations,2020,1-22.
[40]
ZHANG K,ZUO W M,ZHANG L.FFDNet:Toward a fast and flexible solution for CNN-based image denoising[J].IEEE Transactions on Image Processing,2018,27(9):4608-4622.
[41]
GUO S,YAN Z F,ZHANG K,et al.Toward convolutional blind denoising of real photographs[C].Confe-rence on Computer Vision and Pattern Recognition,2019,1712-1722.
[42]
XU J,ZHANG L,ZHANG D,et al.Multi-channel weighted nuclear norm minimization for real color image denoising[C].IEEE Internationa Conference on Computer Vision(ICCV),2017,1105-1113.