Seismic data fault detection based on U-Net deep learning network
YANG Wuyang1,2, YANG Jiarun2,3, CHEN Shuangquan2,3, KUANG Liqin2,3, WANG Enli1,2, ZHOU Chunlei1,2
1. Northwest Branch, Research Institute of Petroleum Exploration & Development, PetroChina, Lanzhou, Gansu 730020, China; 2. CNPC Key Laboratory of Geophysical Prospecting, China University of Petroleum(Beijing), Beijing 102249, China; 3. State Key Laboratory of Petroleum Resources and Prospecting, China University of Petroleum(Beijing), Beijing 102249, China
Abstract：Fault interpretation is one of the key links in seismic data interpretation. With the development of artificial intelligence technology, automatic and rapid fault recognition has become a research hot-spot in the application of machine learning methods in geophysics. At present, intelligent fault recognition is faced with problems, such as difficult model training and the unsatisfactory prediction results of actual data. Therefore, a fault detection method of seismic data based on a U-Net deep learning network is proposed, which combines U-Net and residual module Res-50 in the network structure to construct a new network:ResU-Net. ResU-Net uses the 1×1×1 convolution kernel to process the channel number of feature images. It not only reduces the time complexity but expands the depth of the network based on the original U-Net, effectively improving the operation efficiency and learning ability of the network and identifying faults in a quick and accurate manner. Training and testing of synthetic data sets prove that ResU-Net has less time complexity and solves the problems of fault detection in the case of an irregular data volume by appropriate network input, data expansion, and weighted overlapped boundaries. The application results of actual data show that the ResU-Net training model has strong anti-noise capability, remarkable generalization ability, as well as high prediction accuracy and good continuity of faults.
Bahorich M,Farmer S.3-D seismic discontinuity for faults and stratigraphic features:The coherence cube[J].The Leading Edge,1995,14(10):1053-1058.
[2]
Marfurt K J,Kirlin R L,Farmer S L,et al.3-D seismic attributes using a semblance based coherency algorithm[J].Geophysics,1998,63(4):1150-1165.
[3]
Gersztenkorn A,Marfurt K J.Eigenstructure-based coherence computations as an aid to 3-D structural and stratigraphic mapping[J].Geophysics,1999,64(5):1468-1479.
[4]
Yang L,Gao J,Liu N,et al.A coherence algorithm for seismic data analysis based on the mutual information[J].IEEE Geoscience and Remote Sensing Letters,2019,16(6):967-971.
[5]
仲伟军,姚卫江,贾春明,等.地震多属性断裂识别技术在中拐凸起石炭系中的应用[J].石油地球物理勘探,2017,52(增刊2):135-139.ZHONG Weijun,YAO Weijiang,JIA Chunming,et al.Fault and fracture identification in Carboniferous,Zhongguai Uplift,Junggar Basin with seismic multi-attributes[J].Oil Geophysical Prospecting,2017,52(S2):135-139.
[6]
Andy R.Curvature attributes and their application to 3D interpreted horizons[J].First Break,2001,19(2):85-100.
[7]
吕丙南,陈学华,徐赫,等.空间域加窗二维希尔伯特变换在三维地震资料体边缘检测中的应用[J].石油地球物理勘探,2020,55(3):661-668.LYU Bingnan,CHEN Xuehua,XU He,et al.Application of spatial-windowed 2D Hilbert transform in vo-lumetric edge detection of 3D seismic data[J].Oil Geophysical Prospecting,2020,55(3):661-668.
[8]
姜岩,程顺国,王元波,等.大庆长垣油田断层阴影地震正演模拟及校正方法[J].石油地球物理勘探, 2019,54(2):320-329.JIANG Yan,CHENG Shunguo,WANG Yuanbo,et al.Seismic forward modeling for correction of fault shadow zones in Changyuan Oilfield,Daqing[J].Oil Geophysical Prospecting,2019,54(2):320-329.
[9]
冯连勇,黎莉,裴勇.胜利油田岩性油藏勘探中的地震技术[J].石油地球物理勘探,2006,41(2):211-215.FENG Lianyong,LI Li,PEI Yong.Seismic technology in lithologic reservoir prospecting of Shengli oilfield[J].Oil Geophysical Prospecting,2006,41(2):211-215.
[10]
袁联生.塔里木盆地玉北地区中下奥陶统断溶体识别[J].石油物探,2020,59(4):628-636.YUAN Liansheng.Identifying fault-karst reservoirs in Middle-Lower Ordovician carbonates[J].Geophysical Prospecting for Petroleum,2020,59(4):628-636.
[11]
董守华,石亚丁,汪洋.地震多参数BP人工神经网络自动识别小断层[J].中国矿业大学学报,1997,26(3):16-20.DONG Shouhua,SHI Yading,WANG Yang.Automatic recognition of small fault by BP artificial ner-vous network from multiple seismic parameters[J].Journal of China University of Mining & Technology,1997,26(3):16-20.
[12]
Chehrazi A,Rahimpour-Bonab H,Rezaee M R.Seismic data conditioning and neural network-based attribute selection for enhanced fault detection[J].Petroleum Geoscience,2013,19(2):169-183.
[13]
Huang L,Dong X,Clee T E.A scalable deep learning platform for identifying geologic features from seismic attributes[J].The Leading Edge,2017,36(3):249-256.
[14]
Zhao T,Mukhopadhyay P.A fault-detection workflow using deep learning and image processing[C].SEG Technical Program Expanded Abstracts,2018,37:1966-1970.
[15]
Wu X,Liang L,Shi Y,et al.FaultSeg3D:Using synthetic data sets to train an end-to-end convolutional neural net-work for 3D seismic fault segmentation CNN for 3D fault segmentation[J].Geophysics,2019,84(3):IM35-IM45.
[16]
Liu N,He T,Tian Y,et al.Common azimuth seismic data fault analysis using residual U-Net[J].Interpretation, 2020,8(3):1-41.
[17]
He K,Zhang X,Ren S,et al.Deep residual learning for image recognition[C].Computer Vision and Pattern Recognition,2016,770-778.
[18]
Tingdahl K M,De R M.Semi-automatic detection of faults in 3D seismic data[J].Geophysical Prospecting,2005,53(4):533-542.
[19]
Zhang C,Frogner C,Araya-Polo M,et al.Machine-learning based automated fault detection in seismic traces[C].Extended Abstracts of 76th EAGE Conference and Exhibition, 2014.
Araya-polo M,Dahlke T,Frogner C,et al.Automated fault detection without seismic processing[J].The Leading Edge,2017,36(3):208-214.
[22]
Ronneberger O,Fischer P,Brox T.U-net:Convolutional networks for biomedical image segmentation[C].International Conference on Medical Image Com-puting and Computer-Assisted Intervention,2015,234-241.
[23]
Long J,Shelhamer E,Darrell T.Fully Convolutional Networks for Semantic Segmentation[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,39(4):640-651.
[24]
Xie S,Tu Z.Holistically-nested edge detection[J].International Journal of Computer Vision,2017,125(5):3-18.
