Seismic random noise attenuation based on stationary wavelet transform and deep residual neural network
WU Guoning1, YU Mengmeng1, WANG Junxian1, LIU Guochang2
1. College of Science, China University of Petro-leum(Beijing), Beijing 102249, China; 2. State Key Laboratory of Petroleum Resources and Prospecting, Beijing 102249, China
Abstract:There are many conventional denoising me-thods, but each method is limited by certain assump-tions or conditions. In addition, multiple local extrema may cause the denoising algorithm to converge to a local optimal solution instead of the global one. For this reason, a random noise suppression method based on the stationary wavelet transform and deep residual neural network (WaveResNet) is proposed. It combined the topology structure of the residual neural network (ResNet) with the stationary wavelet transform. The residual module effectively avoids the vanishing gradient or computational consumption caused by the deep network but loss function saturation. In addition, the wavelet transform is an efficient feature extraction method. It can obtain the low-frequency and high-frequency feature information in different directions and learn the characteristics of signal or noise in different regions. First, each picture in the Train400 dataset is rotated by different angles to increase the amount of data in the training set, after which Gaussian noise is added. Then, the one-level stationary Haar wavelet decomposition is performed on each picture to gain a training dataset. The high- and low-frequency information in the wavelet transform domain is extracted through training. On this basis, the wavelet decomposition of the learned noise is subtracted from that of the noisy data, thus achieving the wavelet decomposition of the denoised signal through the direct channel. Finally, the denoised signal is obtained through the inverse stationary wavelet transform. Experiments of synthetic signals and field seismic data show that the proposed method can suppress seismic random noise well, and the signal-to-noise ratio and its peak of the denoised signal are higher than those of conventional methods.
康冶, 于承业, 贾卧, 等.f-x域去噪方法研究[J].石油地球物理勘探, 2003, 38(2):136-138.KANG Ye, YU Chengye, JIA Wo, et al.A study on noise-suppression method in f-x domain[J].Oil Geophysical Prospecting, 2003, 38(2):136-138.
[3]
孔庆丰, 李心友.傅立叶相关系数滤波的实践[J].石油物探, 2001, 40(4):89-93.KONG Qingfeng, LI Xinyou.An application of Fourier correlation coefficient filtering[J].Geophysical Prospecting for Petroleum, 2001, 40(4):89-93.
[4]
蔡剑华, 熊锐.基于频率切片小波变换的时频分析与MT信号去噪[J].石油物探, 2016, 55(6):904-912.CAI Jianhua, XIONG Rui.Magnetotelluric data denosing based on time-frequency analysis of the frequency slice wavelet transform[J].Geophysical Prospecting for Petroleum, 2016, 55(6):904-912.
[5]
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.
[6]
张之涵, 孙成禹, 姚永强, 等.三维曲波变换在地震资料去噪处理中的应用研究[J].石油物探, 2014, 53(4):421-430.ZHANG Zhihan, SUN Chengyu, YAO Yongqiang, et al.Research on the application of 3D Curvelet transform to seismic data denoising[J].Geophysical Prospecting for Petroleum, 2014, 53(4):421-430.
[7]
张入化, 黄建平, 国运东, 等.基于Seislet域分数阶阈值去噪算法的地震资料去噪[J].石油物探, 2020, 59(1):40-50.ZHANG Ruhua, HUANG Jianping, GUO Yundong, et al.Fractional threshold denoising algorithm in Seislet domain for seismic data denoising[J].Geophysical Prospecting for Petroleum, 2020, 59(1):40-50.
[8]
FOMEL S and LIU Y.Seislet transform and seislet frame[J].Geophysics, 2010, 75(3):V25-V38.
[9]
尚平萍, 李鹏, 杨安琪, 等.基于CEEMDAN的地震信号高分辨率时频分析方法[J].石油物探, 2019, 58(4):547-554.SHANG Pingping, LI Peng, YANG Anqi, et al.Seismic high-resolution time-frequency analysis based on CEEMDAN[J].Geophysical Prospecting for Petroleum, 2019, 58(4):547-554.
[10]
HUANG N E, SHEN Z, LONG S R, et al.The empi-rical mode decomposition and the Hilbert spectrum for nonlinear non-stationary time series analysis[J].Proceeding of the Royal Society of A:Mathematical, Physical and Engineering Sciences, 1998, 454(1971):903-995.
[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]
DRAGOMIRETSKIY K, ZOSSO D.Variational mode decomposition[J].IEEE Transactions on Signal Processing, 2014, 62(3):531-544.
[13]
方江雄, 温志平, 顾华奇, 等.基于变分模态分解的地震随机噪声压制方法[J].石油地球物理勘探, 2019, 54(4):757-767.FANG Jiangxiong, WEN Zhiping, GU Huaqi, et al.Seismic random noise attenuation based on variational mode decomposition[J].Oil Geophysical Prospecting, 2019, 54(4):757-767.
[14]
李江.基于奇异值分解的角度域去噪方法[J].石油物探, 2019, 58(3):427-432.LI Jiang.Seismic denoising in the angle domain based on singular value decomposition[J].Geophysical Prospecting for Petroleum, 2019, 58(3):427-432.
[15]
胡永泉, 黄建波, 田志华, 等.基于单道SVD和振幅比的地面微地震资料去噪方法[J].石油物探, 2019, 58(1):43-52.HU Yongquan, HUANG Jianbo, TIAN Zhihua, et al.Ground microseismic data denoising based on single-channel singular value decomposition and amplitude ratio[J].Geophysical Prospecting for Petroleum, 2019, 58(1):43-52.
[16]
毛海波, 马俊彦, 王晓涛, 等.基于自适应字典学习的可控震源数据谐波噪声压制方法[J].石油物探, 2020, 59(5):725-735.MAO Haibo, MA Junyan, WANG Xiaotao, et al.Harmonic noise suppression of vibroseis data based on adaptive dictionary learning[J].Geophysical Prospecting for Petroleum, 2020, 59(5):725-735.
[17]
HE K M, ZHANG X Y, REN S Q, et al.Deep residual learning for image recognition[C].2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, doi:10.1109/CVPR.2016.90.
[18]
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.
[19]
SAAD O M, CHEN Y K.Deep denoising autoencoder for seismic random noise attenuation[J].Geophysics, 2020, 85(4):V367-V376.
[20]
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 learning convolutional neural networks for random noise attenuation in seismic data[J].Geophysical Prospecting for Petroleum, 2018, 57(6):862-869.
[21]
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.
[22]
YU S W, MA J W, and WANG W L.Deep learning for denoising[J].Geophysics, 2019, 84(6):V333-V350.
[23]
ZHOU Y Y and WU G N.Unsupervised machine lear-ning for waveform extraction in microseismic denoising[J].Journal of Applied Geophysics, 2020, doi:10.1016/j.jappgeo.2019.103879.
[24]
张岩, 李新月, 王斌, 等.基于联合深度学习的地震数据随机噪声压制[J].石油地球物理勘探, 2021, 56(1):9-25, 56.ZHANG Yan, LI Xinyue, WANG Bin, et al.Random noise suppression of seismic data based on joint deep learning[J].Oil Geophysical Prospecting, 2021, 56(1):9-25, 56.
[25]
唐杰, 孟涛, 张文征, 等.利用基于深度学习的过完备字典信号稀疏表示算法压制地震随机噪声[J].石油地球物理勘探, 2020, 55(6):1202-1209.TANG Jie, MENG Tao, ZHANG Wenzheng, et al.Suppressing seismic random noise based on Deep-KSVD[J].Oil Geophysical Prospecting, 2020, 55(6):1202-1209.
[26]
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.
[27]
KANG E, MIN J H, YE J C.Wavelet domain residual network(WaveResNet) for low-dose X-ray CT reconstruction[EB/OL].[2018-03-28].https://arxiv.org/abs/1707.09938v2.
[28]
GU J and YE J C.Multi-scale wavelet domain residual learning for limited-angle CT reconstruction[EB/OL].[2017-03-04].https://arxiv.org/abs/1703. 01382v1.
[29]
ZHANG F, XU Z C, CHEN W, et al.An image compression method for video surveillance system in underground mines based on residual networks and discrete wavelet transform[J].Electronics, 2019, doi:10.3390/electronics8121559.
[30]
LIU Pengju, ZHANG H Z, ZHANG K, et al.Multi-level wavelet-CNN for image restoration[C].2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops(CVPRW), 2018, doi:10.1109/CVPRW.2018.00121.
[31]
FOWLER J E.The redundant discrete wavelet transform and additive noise[J].IEEE Signal Processing Letters, 2005, 12(9):629-632.
[32]
HE K M, ZHANG X Y, REN S Q, et al.Deep resi-dual learning for image recognition[C].2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, doi:10.1109/CVPR.2016.90.