Seismic resolution improvement method based on dual-attention U-Net network
LI Xuegui1,2,3, ZHOU Yingjie2, DONG Hongli1, WU Jun4, XU Gang5, WANG Ruyi6
1. Institute of Artificial Energy Research, Northeast Petroleum University, Daqing, Heilongjiang 163318, China; 2. School of Computer and Information Technology, Northeast Petroleum University, Daqing, Heilongjiang 163318, China; 3. Heilongjiang Provincial Key Laboratory of Big Data and Intelligent Analysis of Petroleum, Daqing, Heilongjiang 163318, China; 4. Exploration and Development Research Institute of Daqing Oilfield, Daqing, Heilongjiang 163712, China; 5. New Resources Geophysical Exploration Division, BGP Inc., Zhuozhou, Hebei 072751, China; 6. CNPC Research Institute of Petroleum Engineering Co., Ltd, Beijing 102206, China
Abstract:Traditional methods to improve the resolution of seismic data, such as deconvolution and Q compensation, are limited by the assumption that the wavelet is the minimum phase, and the reflection coefficient is white noise, and they need to calculate complex parameters, which brings inconvenience to practical application. The deep learning method uses the data-driven method to adaptively depict the relationship between input and target and has excellent self-learning ability. However, the current method of improving seismic data resolution based on deep learning does not fully utilize attention information. Therefore, a method of improving seismic data resolution based on dual attention U-Net network is proposed. First of all, the improved channel attention module, spatial attention module, and cascade residual module are added to the original U-Net network, which can not only quickly learn the mapping relationship between high- and low-resolution data but also reasonably allocate the weight of different channels and spaces and make full use of the correlation between data; then, the combination of L1 loss and loss of multi-scale-structural similarity index measurement is used as the loss function to improve the sensitivity of the model to local information changes and facilitate the recovery of detail information. The test results of simulated data and actual data show that this method improves the main frequency and frequency band width of seismic data, makes the event axis clearer, enriches the detailed texture information, and effectively improves the resolution of seismic data.
钱勇先.重采样法高分辨率处理[J].石油地球物理勘探,1994,29(4):498-503.QIAN Yongxian.High-resolution data processing by resampling method[J].Oil Geophysical Prospecting,1994,29(4):498-503.
[2]
CHEN S,WANG Y.Seismic resolution enhancement by frequency-dependent wavelet scaling[J].IEEE Geo-science and Remote Sensing Letters,2018,15(5):654-658.
[3]
刘瑞,何文章,吴爱弟,等.基于多小波变换提高地震数据的分辨率[J].数学的实践与认识,2011,41(20):43-47.LIU Rui,HE Wenzhang,WU Aidi,et al.Improving the resolution ratio of seismic data by multiwavelets transformation[J].Mathematics in Practice and Theory,2011,41(20):43-47.
[4]
WANG Q,GAO J,LIU N,et al.High-resolution seismic time-frequency analysis using the synchrosquee-zing generalized S-Transform[J].IEEE Geo-science and Remote Sensing Letters,2018,15(3):374-378.
[5]
郭爱华,路鹏飞,余波,等.利用Shearlet变换提高叠后地震资料分辨率[J].石油地球物理勘探,2021,56(5):992-1000.GUO Aihua,LU Pengfei,YU Bo,et al.Use shearlet transform to improve the resolution of the stacked seismic data[J].Oil Geophysical Prospecting,2021,56(5):992-1000.
[6]
李思源,徐天吉.基于Wigner-Ville分布与Chrip-Z变换的高分辨时频分析方法[J].石油地球物理勘探,2022,57(1):168-175,211.LI Siyuan,XU Tianji.High-resolution time-frequency analysis method based on Wigner-Ville distribution and Chrip-Z transformation[J].Oil Geophysical Prospecting,2022,57(1):168-175,211.
[7]
WINARDHI S,PRANOWO W.Multi-ricker spectral modeling in the S-transform domain for enhancing vertical resolution of seismic reflection data[J].Indonesian Journal on Geoscience,2019,6(3):223-233.
[8]
LIU G,LI C,RAO Y,et al.Oriented pre-stack inverse Q filtering for resolution enhancements of seismic data[J].Geophysical Journal International,2020,223(1):488-501.
[9]
杜鑫,张广智,王若.一种提高浅层地震资料分辨率的反褶积组合方法应用效果[C].中国石油学会2021年物探技术研讨会论文集,2021,304-307.DU Xin,ZHANG Guangzhi,WANG Ruo.Application effect of a deconvolution combination method to improve the resolution of shallow seismic data[C].Proceedings of 2021 Geophysical Exploration Technology Seminar of China Petroleum Society,2021,304-307.
[10]
YUAN Y,LI Y,ZHOU S.Multichannel statistical broadband wavelet deconvolution for improving resolution of seismic signals[J].IEEE Transactions on Geoscience and Remote Sensing,2021,59(2):1772-1783.
[11]
SAJID M,GHAZALI A R.Nonstationary differential resolution:An algorithm to improve seismic resolution[J].Geophysics,2018,83(3):V149-V156.
[12]
吴大奎,吴宗蔚,伍翊嘉.叠后地震资料井控高分辨率处理新方法[J].天然气工业,2019,39(11):36-43.WU Dakui,WU Zongwei,WU Yijia.A new method of well-control high resolution processing for post-stack seismic data[J].Natural Gas Industry,2019,39(11):36-43.
[13]
ZHANG G,DUAN J,LI Y,et al.Adaptive time-resam-pled high-resolution synchrosqueezing transform and its application in seismic data[J].IEEE Transactions on Geoscience and Remote Sensing,2020,58(9):6691-6698.
[14]
刘汉卿,罗明,孙辉,等.应用匹配追踪算法的高分辨率处理方法[J].石油地球物理勘探,2022,57(6):1325-1331.LIU Hanqing,LUO Ming,SUN Hui,et al.High-resolution processing method based on matching pursuit algorithm[J].Oil Geophysical Prospecting,2022,57(6):1325-1331.
[15]
吴大奎,李亚林,伍志明,等.地震、测井资料联合反演的神经网络算法研究[J].天然气工业,2004,24(3):55-57.WU Dakui,LI Yalin,WU Zhiming,et al.Research on the neural network algorithm of joint inversion of seismic and log data[J].Natural Gas Industry,2004,24(3):55-57.
[16]
宋维琪,赵万金,吴华,等.利用多分辨率小波网络进行地震资料反演[J].地震地质,2005,27(1):98-104.SONG Weiqi,ZHAO Wanjin,WU Hua,et al.Inversion of seismic data by using multi-resolution wavelet networks[J].Seismology and Geology,2005,27(1):98-104.
[17]
王姣,李红梅,李振春,等.基于GRNN振幅谱估计的井控提高地震分辨率技术[J].石油学报,2015,36(6):715-723.WANG Jiao,LI Hongmei,LI Zhenchun,et al.Well-controlled seismic resolution enhancement technology based on GRNN amplitude spectrum estimation[J].Acta Petrolei Sinica,2015,36(6):715-723.
[18]
蔡涵鹏,李会强,秦情,等.基于生成对抗网络地震数据超分辨率重建方法[C].中国石油学会2019年物探技术研讨会论文集,2019,1547-1550.CAI Hanpeng,LI Huiqiang,QIN Qing,et al.Super resolution reconstruction method of seismic data based on generation countermeasure network[C].Proceedings of 2019 Geophysical Exploration Technology Seminar of China Petroleum Society,2019,1547-1550.
[19]
李祺鑫,罗亚能.生成对抗网络高分辨率地震反演[C].中国石油学会2019年物探技术研讨会论文集,2019,1475-1478.LI Qixin,LUO Yaneng.Generate countermeasure network high-resolution seismic inversion[C].Procee-dings of 2019 Geophysical Exploration Technology Seminar of China Petroleum Society,2019,1475-1478.
[20]
RONNEBERGER O,FISCHER P,BROX T.U-Net:convolutional networks for biomedical image segmentation[C].Medical Image Computing and Computer-Assisted Intervention - MICCAI,2015,234-241.
[21]
孙永壮,黄鋆,俞伟哲,等.基于U-Net网络的端到端地震高分辨率处理技术[J].地球物理学进展,2021,36(3):1297-1305.SUN Yongzhuang,HUANG Yun,YU Weizhe,et al.End to end seismic high resolution processing technology based on U-Net network[J].Progress in Geophysics,2021,36(3):1297-1305.
[22]
LI J,WU X,HU Z.Deep learning for simultaneous seismic image super-resolution and denoising[J].IEEE Transactions on Geoscience and Remote Sen-sing,2022,60:1-11.
[23]
陈志明,刘洋,张浩然.基于U-Net网络的地震资料高分辨率处理[C].中国石油学会2021年物探技术研讨会论文集,2021,988-991.CHEN Zhiming,LIU Yang,ZHANG Haoran.High resolution processing of seismic data based on U-Net network[C].Proceedings of 2021 Geophysical Exploration Technology Seminar of China Petroleum Society,2021,988-991.
[24]
BAHDANAU D,CHO K H,BENGIO Y.Neural machine translation by jointly learning to align and translate[C].3rd International Conference on Lear-ning Representations,(ICLR) 2015,San Diego,2015.
[25]
LI L,ZOU Y,WANG B,et al.Terahertz image super-resolution using an improved attention U-Net[C].Computational Imaging VI,2021,69-73.
[26]
CHEN R,ZHANG H,LIU J.Multi-attention augmented network for single image super-resolution[J].Pattern Recognition,2022,122:108349.
[27]
AI S,KWON J.Extreme low-light image enhancement for surveillance cameras using attention U-Net[J].Sensors,2020,20(2):495.
[28]
武瑞沁,陈雪勤,俞杰,等.结合注意力机制的改进U-Net网络在端到端语音增强中的应用[J].声学学报,2022,47(2):266-275.WU Ruiqin,CHEN Xueqin,YU Jie,et al.Application of improved U-Net network with attention mechanism in end-to-end speech enhancement[J].Acta Acustica,2022,47(2):266-275.
[29]
殷晓航,王永才,李德英.基于U-Net结构改进的医学影像分割技术综述[J].软件学报,2021,32(2):519-550.YIN Xiaohang,WANG Yongcai,LI Deying.Suvery of medical image segmentation technology based on U-Net structure improvement[J].Journal of Software,2021,32(2):519-550.
[30]
谢学立,李传祥,杨小冈,等.双注意力循环卷积显著性目标检测算法[J].光学学报,2019,39(9):260-270.XIE Xueli,LI Chuanxiang,YANG Xiaogang,et al.Salient object detection algorithm based on dual-attention recurrent convolution[J].Acta Optica Sinica,2019,39(9):260-270.
[31]
LIM B,SON S,KIM H,et al.Enhanced deep residual networks for single image super-resolution[C].2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW),2017,1132-1140.
[32]
SHI W,CABALLERO J,HUSZÁR F,et al.Real-time single image and video super-resolution using an efficient sub-pixel convolutional neural network[C].2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2016,1874-1883.
[33]
周文辉,石敏,朱登明,等.基于残差注意力网络的地震数据超分辨率方法[J].计算机科学,2021,48(8):24-31.ZHOU Wenhui,SHI Min,ZHU Dengming,et al.Seismic data super-resolution method based on residual attention network[J].Computer Science,2021,48(8):24-31.
[34]
WU X,GENG Z,SHI Y,et al.Building realistic structure models to train convolutional neural networks for seismic structural interpretation[J].Geophysics,2019,85(4):WA27-WA39.
[35]
陈国民,沈章洪,鲍祥生,等.确定合成地震记录子波主频的方法研究[J].石油天然气学报,2009,31(5):55-60.CHEN Guomin,SHEN Zhanghong,BAO Xiang-sheng,et al.Method to determine the dominant fre-quency of the synthetic seismogram[J].Journal of Oil and Gas Technology,2009,31(5):55-60.
[36]
楼鑫杰,李小薪,刘志勇.基于反馈机制的图像超分辨率重建算法[J].计算机工程,2022,48(2):261-267.LOU Xinjie,LI Xiaoxin,LIU Zhiyong.Super-resolution image reconstruction algorithm based on feedback mechanism[J].Computer Engineering,2022,48(2):261-267.
[37]
江明,肖庆生,易见兵,等.自适应残差注意力的轻量级图像超分辨率重建模型[J].激光与光电子学进展,2022,59(16):133-141.JIANG Ming,XIAO Qingsheng,YI Jianbing,et al.Light weight super-resolution image-reconstruction model with adaptive residual attention[J].Laser & Optoelectronics Progress,2022,59(16):133-141.
[38]
WANG Z,BOVIK A C,SHEIKH H R,et al.Image quality assessment:from error visibility to structural similarity[J].IEEE Transactions on Image Proces-sing,2004,13(4):600-612.
[39]
ZHAO H,GALLO O,FROSIO I,et al.Loss functions for image restoration with neural networks[J].IEEE Transactions on Computational Imaging,2017,3(1):47-57.
[40]
KINGMA D P,BA J.Adam:a method for stochastic optimization[C].Contribution to International Conference on Learning Representations,San Diego,2014.