1. School of Geosciences, China University of Petroleum(East China), Qingdao, Shandong 266580, China; 2. Research Institute of Exploration and Development, SINOPEC Shengli Oilfied Company, Dongying, Shangdong 257105, China
Abstract：Curvelet transform denoising causes the events to be distorted and interferes with the effective signals in discontinuous areas such as fault zones. The algorithm of overcomplete dictionaries for sparse representation (K-SVD) requires manual and repeated adjustment of parameters to improve the denoising effect. After comprehensively considering the advantages of deep learning network and sparse representation,we combined the K-SVD denoising algorithm with the deep learning network,and proposed the random noise suppression method based on Deep-KSVD.In order to make the network have the ability to learn parameters,the OMP algorithm is replaced by an equivalent learnable alternative in the tracking phase.The calculation process includes decomposing seismic data into overlapped data blocks,de-noising each data block by proper tracking,and reconstructing the whole data by weighting the denoised data blocks.The denoising process includes three parts:sparse coding,estimation of regularized coefficient and reconstruction of data block.The test results on model data and actual data show that after training a Deep-KSVD network,for given noisy data,it can adaptively attenuate the seismic noises without further adjusting parameters while protecting the effective information of discontinuity and the characteristics of data structure.Compared with the K-SVD denoising method,the Deep-KSVD denoising method provides better effect of noise suppression and can improve the signal-to-noise ratio of full-band data.
陈颖频,彭真明,李美惠,等.基于交叠组稀疏广义全变分的地震信号随机噪声衰减[J].石油地球物理勘探,2019,54(1):24-35,44.CHEN Yingpin,PENG Zhenming,LI Meihui,et al.Random noise suppression based on the improved total generalized variation with overlapping group sparsity[J].Oil Geophysical Prospecting,2019,54(1):24-35,44.
[2]
唐杰,张文征,梁雨薇,等.自适应数据驱动的紧框架微地震数据随机噪声压制[J].石油地球物理勘探,2019,54(5):954-961.TANG Jie,ZHANG Wenzheng,LIANG Yuwei,et al.A random-noise suppression approach with self-adaptive data-driven tight frame for microseismic data[J].Oil Geophysical Prospecting,2019,54(5):954-961.
[3]
唐杰,戚瑞轩,张文征,等.基于自相似块匹配的地震数据信噪分离方法研究[J].石油物探,2020,59(2):198-207.TANG Jie,QI Ruixuan,ZHANG Wenzheng,et al.Seismic data denoising based on self-similarity block matching[J].Geophysical Prospecting for Petroleum,2020,59(2):198-207.
[4]
曹静杰,杨志权,杨勇,等.一种基于曲波变换的自适应地震随机噪声消除方法[J].石油物探,2018,57(1):72-78.CAO Jingjie,YANG Zhiquan,YANG Yong,et al.An adaptive seismic random noise elimination method based on Curvelet transform[J].Geophysical Prospecting for Petroleum,2018,57(1):72-78.
[5]
张华,刁塑,温建亮,等.应用二维非均匀曲波变换压制地震随机噪声[J].石油地球物理勘探,2019,54(1):16-23.ZHANG Hua,DIAO Su,WEN Jianliang,et al.A random noise suppression with 2D non-uniform Curvelet transform[J].Oil Geophysical Prospecting,2019,54(1):16-23.
[6]
孙成禹,刁俊才,李文静.基于曲波噪声估计的三维块匹配地震资料去噪[J].石油地球物理勘探,2019,54(6):1188-1194.SUN Chengyu,DIAO Juncai,LI Wenjing.3D Block matching seismic data denoising based on Curvelet noise estimation[J].Oil Geophysical Prospecting,2019,54(6):1188-1194.
[7]
薛林,程浩,巩恩普,等.Shearle域自适应阈值地震数据随机噪声压制[J].石油地球物理勘探,2020,55(2):282-291.XUE Lin,CHENG Hao,GONG Enpu,et al.Random noise suppression using adaptive threshold in Shearlet domain[J].Oil Geophysical Prospecting,2020,55(2):282-291.
[8]
童思友,高航,刘锐,等.基于Shearlet变换的自适应地震资料随机噪声压制[J].石油地球物理勘探,2019,54(4):744-750.TONG Siyou,GAO Hang,LIU Rui,et al.Seismic random noise adaptive suppression based on the Shearlet transform[J].Oil Geophysical Prospecting,2019,54(4):744-750.
[9]
李勇,张益明,雷钦,等.模型约束下的在线字典学习地震弱信号去噪方法[J].地球物理学报,2019,62(1):411-420.LI Yong,ZHANG Yiming,LEI Qin,et al.Online dictionary learning seismic weak signal denoising method under model constraints[J].Chinese Journal of Geophysics,2019,62(1):411-420.
[10]
张良,韩立国,方金伟,等.双稀疏字典和FISTA的地震数据去噪[J].地球物理学报,2019,62(7):2671-2683.ZHANG Liang,HAN Liguo,FANG Jinwei,et al.Seismic data denoising via double sparsity dictionary and fast iterative shrinkage-thresholding algorithm[J].Chinese Journal of Geophysics,2019,62(7):2671-2683.
[11]
乐友喜,杨涛,曾贤德.CEEMD与KSVD字典训练相结合的去噪方法[J].石油地球物理勘探,2019,54(4):729-736.YUE Youxi,YANG Tao,ZENG Xiande.Seismic denoising with CEEMD and KSVD dictionary combined training[J].Oil Geophysical Prospecting,2019,54(4):729-736.
[12]
Hinton G E,Osindero S,Teh Y W.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527-1554.
[13]
Vincent P,Larochelle H,Lajoie I,et al.Stacked denoising autoencoders:Learning useful representations in a deep network with a local denoising criterion[J].Journal of Machine Learning Research,2010,11(12):3371-3408.
[14]
Gu J X,Wang Z H,Jason K,et al.Recent advances in convolutional neural networks[J].Pattern Recognition,2018,77(5):354-377.
[15]
王琪琪,汤井田,张良,等.利用多层感知机的地震数据去噪[J].石油地球物理勘探,2020,55(2):272-281.WANG Qiqi,TANG Jingtian,ZHANG Liang,et al.Seismic data denoising based on multi-layer perceptron[J].Oil Geophysical Prospecting,2020,55(2):272-281.
[16]
Zhang K,Zuo W,Chen Y,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.
[17]
Wu H,Zhang B,Lin T F,et al.White noise attenuation of seismic trace by integrating variational mode decomposition with convolutional neural network[J].Geophysics,2019,84(5):V307-V317.
[18]
Yu S W,Ma J W,Wang W L.Deep learning for denoising[J].Geophysics,2019,84(6):V333-V350.
[19]
Jia Y N,Ma J W.What can machine learning do for seismic data processing?An interpolation application[J].Geophysics,2017,82(3):V163-V177.
Aharon M,Elad M,Bruckstein A.K-SVD:An algorithm for designing overcomplete dictionaries for sparse representation[J].IEEE Transactions on Signal Processing,2006,54(11):4311-4322.
[22]
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.
[23]
Chen S S,Donoho D L,Saunders M A.Atomic decomposition by basis pursuit[J].SIAM Review,2001,43(1):129-159.
[24]
Romano Y,Aberdam A,Sulam J,et al.Adversarial noise attacks of deep learning architectures:stability analysis via sparse-modeled signals[J].Journal of Mathematical Imaging and Vision,2019,62(5):313-327.
[25]
Donoho D L,Elad M.Optimally sparse representation in general (nonorthogonal) dictionaries via L1 mini-mization[J].National Academy of Sciences,2003,100(5):2197-2202.
[26]
Aberdam A,Sulam J,Elad M.Multi layer sparse coding:The holistic way[J].SIAM Journal on Mathematics of Data Science,2019,1(1):46-77.
[27]
Rey-Otero I,Sulam J,Elad M.Variations on the CSC model[J].IEEE Transactions on Signal Processing,2020,68(1):519-528.
[28]
Mairal J,Sapiro G,Elad M.Learning multiscale sparse representations for image and video restoration[J].Multiscale Modeling & Simulation,2008,7(1):214-241.
[29]
Mao X J,Shen C,Yang Y B.Image restoration using very deep convolutional encoder-decoder networks with symmetric skip connections[C].Proceedings of the 30th International Conference on Neural Information Processing Systems,2016,2810-2818.
[30]
Papyan V,Romano Y,Elad M.Convolutional neural networks analyzed via convolutional sparse coding[J].The Journal of Machine Learning Research,2017,18(1):2887-2938.
[31]
Ophir B,Lustig M,Elad M.Multi-scale dictionary learning using wavelets[J].IEEE Journal of Selected Topics in Signal Processing,2011,5(5):1014-1024.
[32]
Papyan V,Elad M.Multi-scale patch-based image restoration[J].IEEE Transactions on Image Processing,2016,25(1):249-261.
