An improved neural-network cascade-correlation algorithm and its application in seismic first break picking
Song Jianguo1, Li Fuzhen2, Xu Weixiu3, Li Zhe1
1. School of Geosciences, China University of Petroleum(East China), Qingdao, Shandong 266580, China; 2. Dagang Branch, BGP Inc., CNPC, Tianjin 300280, China; 3. Shengli Branch, Geophysical Company Ltd., SINOPEC, Dongying, Shandong 257600, China
Abstract:To overcome existing problems of BP neural network,we introduce the cascade-correlation algorithm into the neural network construction.The proposed algorithm has a faster convergence than BP algorithm,and it can decide its own network architecture based on problems to be solved.That means it can expand network topology to learn new samples.The initial network of the standard cascade-correlation algorithm has only an input layer and an output layer,while the improved algorithm starts with an appropriate BP network architecture (including hidden layers).In addition,in order to prevent weight-ill growth,a regularization term is added to objective functions in candidate hidden units training to decay weights.Simulation experiments demonstrate that the improved cascade-correlation algorithm has faster convergence and stronger generalization ability.Cross-plots of five attributes such as instantaneous-intensity ratio,amplitude,frequency,curve-length ratio,and adjacent-trace correlation,show that first breaks peaked by the proposed algorithm can be easily and reliably discriminated.The proposed algorithm achieves good performance in the first break picking on real data.
宋建国, 李赋真, 徐维秀, 李哲. 改进的神经网络级联相关算法及其在初至拾取中的应用[J]. 石油地球物理勘探, 2018, 53(1): 8-16.
Song Jianguo, Li Fuzhen, Xu Weixiu, Li Zhe. An improved neural-network cascade-correlation algorithm and its application in seismic first break picking. Oil Geophysical Prospecting, 2018, 53(1): 8-16.
曾庆才,曾同生,欧阳永林等.复杂山地层析反演静校正新方法及应用.石油地球物理勘探,2017,52(3):418-425,441.Zeng Qingcai,Zeng Tongsheng,Ouyang Yonglin et al.A new tomographic inversion for static corrections in complex mountain areas.OGP,2017,52(3):418-425,441.
[2]
谢春,刘玉柱,董良国等.基于声波方程的有限频伴随状态法初至波旅行时层析.石油地球物理勘探,2015,50(2):274-282.Xie Chun,Liu Yuzhu,Dong Liangguo et al.First arrival tomography with finite frequency adjoint state method based on acoustic wave equation.OGP,2015,50(2):274-282.
[3]
杨国权,丁鹏程,李振春等.应用线性程函方程和整形正则化的三维初至波旅行时层析.石油地球物理勘探,2017,52(2):264-272.Yang Guoquan,Ding Pengcheng,Li Zhenchun et al.3D first arrival traveltime tomography with linearized eikonal equation and shaping regularization.OGP,2017,52(2):264-272.
[4]
左国平,王彦春,隋荣亮.利用能量比法拾取地震初至的一种改进方法.石油物探,2004,43(4):345-347.Zuo Guoping,Wang Yanchun,Sui Rongliang.An improved method for first arrival pickup using energy ratio.GPP,2004, 43(4):345-347.
[5]
张伟,王彦春,李洪臣等.地震道瞬时强度比法拾取初至波.地球物理学进展,2009,24(1):201-204.Zhang Wei,Wang Yanchun,Li Hongchen et al.Seismic first arrival pickup using instantaneous intensity ratio.Progress in Geophysics,2009,24(1):201-204.
[6]
牛沛琛,张建中.基于数字图像处理的地震波初至时间检测.计算机与现代化,2007,(4):27-30.Niu Peichen,Zhang Jianzhong.First arrival picking based on image processing methods.Computer and Modernization,2007,(4):27-30.
[7]
曹茂森,任青文,万林梅等.Length分形维算法拾取地震波初至.石油地球物理勘探,2004,39(5):509-514.Cao Maosen,Ren Qingwen,Wang Linmei et al.Length fractal dimension method for picking siemic first break.OGP,2004,39(5):509-514.
[8]
裴正林,余钦范.基于小波变换和BP神经网络的地震波初至拾取方法.勘察科学技术,1999,(4):61-64.Pei Zhenglin,Yu Qinfan.A wavelet transform and BP neural network-based algorithm for detecting first arrivals on seismic waves.Site Investigation Science and Technology,1999,(4):61-64.
[9]
Murat M E and Rudman A J.Automated first arrival picking:A neural network approach.Geophysical Prospecting,1992,40(6):587-604.
[10]
王金峰,罗省贤.BP神经网络的改进及其在初至波拾取中的应用.物探化探计算技术,2006,28(1):14-17.Wang Jinfeng,Luo Shengxian.The improved BP neural network and its application in seismic first breaks picking.Computing Techniques for Geophysical and Geochemical Exploration,2006,28(1):14-17.
[11]
Kusuma T and Brown M M.Cascade-correlation learning architecture for first-break picking and automated trace editing.SEG Technical Program Expanded Abstractsx,1992,11:1136-1139.
[12]
Fahlman S E,Lebiere C.The cascade-correlation learning architecture.Advances in Neural Information Processing Systems,1990,524-532.
[13]
Baluja S,Fahlman S E.Reducing Network Depth in the Cascade-Correlation Learning Architecture.School of Computer Science,Carnegie Mellon University,1994,1-11.
[14]
杨钟瑾,史忠科.快速自底向上构造神经网络的方法.数学实践与认识,2004,34(9):114-118.Yang Zhongjin,Shi Zhongke.Fast approach for cascade-correlation learning.Mathematics in Practice and Theory,2004,34(9):114-118.
[15]
杨慧中,王伟娜,丁锋.神经网络的两种结构优化算法研究.信息与控制,2006,35(6):700-705.Yang Huizhong,Wang Weina,Ding Feng.Two structure optimization algorithms for neural networks.Information and Control,2006,35(6):700-705.
[16]
王治国,尹成,唐何兵等.沉积模式约束的地震多属性水下扇岩相划分.石油地球物理勘探,2015,50(2):357-362.Wang Zhiguo,Yin Cheng,Tang Hebing et al.Submarine fan lithofacies identification with depositional model and seismic attributes.OGP,2015,50(2):357-362.
[17]
Hart D I.Improving the reliability of first-break picking with neural networks.SEG Technical Program Expanded Abstracts,1996,15:1662-1665.
[18]
Wu Q,Nakayama K.Avoiding weight-ill growth:cascade correlation algorithm with local regularization.Neural Networks,1997,3,1954-1959.
[19]
Fahlman S E.Fast learning variations on back-propa-gation:An empirical study.In Proceedings of the 1988 Connectionist Models Summer School,1989,38-51.
[20]
顾汉明,周鸿秋,张学强.初至时间的自动拾取.物探与化探.1992,16(2):120-129.Gu Hanming,Zhou Hongqiu,Zhang Xueqiang.Automatic pick of first arrival time.Geophysical and Geochemical Exploration,1992,16(2):120-129.
[21]
印兴耀,韩文功,李振春等.地震技术新进展(第一版).山东东营:中国石油大学出版社,2006,58-75.Yin Xingyao,Han Wengong,Li Zhenchun et al.New Progress in Seismic Technology (First Version).The Publishing House of China University of Petroleum,Dongying,Shandong,2006,58-75.
[22]
McCormack M D,Zaucha D E and Dushek D W.First-break refraction event picking and seismic data trace editing using neural networks.Geophysics,1993,58(1):67-78.