基于加权<i>K</i>均值聚类的多属性初至拾取方法

周竹生; 曾维祖; 刘思琴; 陈文样

doi:10.11939/jass.20190107

基于加权K均值聚类的多属性初至拾取方法

1.
中国长沙 410012 中南大学地球科学与信息物理学院
2.
中国长沙 410083 中南大学有色金属成矿预测与地质环境监测教育部重点实验室

详细信息

通讯作者:
周竹生: e-mail：geophys@126.com

中图分类号: P315.3⁺1
计量
- 文章访问数: 914
- HTML全文浏览量: 575
- PDF下载量: 52
出版历程
- 收稿日期: 2019-06-13
- 修回日期: 2019-10-23
- 录用日期: 2019-10-23
- 网络出版日期: 2020-05-21
- 刊出日期: 2020-05-20

First arrival picking method by seismic multi-attribute based on weighted K-means clustering algorithm

1.
School of Geosciences and Info-Physics，Central South University，Changsha 410012，China
2.
Key Laboratory of Metallogenic Prediction of Nonferrous Metals and Geological Environment Monitoring （Central South University）， Ministry of Education，Changsha 410083，China

摘要

摘要:
为提高初至拾取方法的准确性和自适应能力，将变异系数加权K均值聚类算法引入初至拾取中。首先提取均方根振幅、相邻道相关性、线积分、振幅谱主频等多种地震属性；然后针对地震属性进行加权K均值聚类，自动识别初至所在时窗；最后结合相位校正法，实现时窗内初至波起跳时间的拾取。在此基础上通过实际数据测试，并与长短时窗能量比法、反向传播神经网络方法对比，验证了本文方法的有效性与可行性。结果表明，基于加权K均值聚类的多属性初至拾取方法能较快速、准确地拾取低信噪比数据的初至，并且无需人为判断时窗，从而提高了拾取的自适应能力。
- 地震属性 /
- 加权K均值聚类 /
- 初至拾取 /
- 初至波
Abstract:
For the purpose of improving the accuracy and automation of first arrival picking method, the weighted K-means clustering algorithm is introduced. Firstly, various seismicattributes such as root-mean-squares amplitude, adjacent trace correlation, line integral and dominant frequency of amplitude spectrum are extracted. Then, weighted K-means clustering is performed for seismic attributes to identify the first arrival time window automatically.Finally, combined with phase correction method, this method is applied to realize the pickup of the first arrival time in the time window. The validity and feasibility of the proposed method, which is compared with STA/LTA and BP neural network, are verified by theoretical and practical data tests. The results suggest that the multi-attribute first arrival picking method based on weighted K-means clustering can pick up the first arrival of seismic data with low signal-to-noise ratios quickly and accurately, and enhance the automation of arrival time picking without the artificial identification of time window.
- seismic attribute /
- weighted-K-means clustering /
- first-arrival picking /
- first arrival wave

HTML全文

图 1 K均值算法流程图

Figure 1. Flowchart of K-means algorithm

下载: 全尺寸图片幻灯片

图 2 理论速度模型

Figure 2. Theoretical velocity model

下载: 全尺寸图片幻灯片

图 3 理论模型正演记录

Figure 3. Seismic records for theoretical model

下载: 全尺寸图片幻灯片

图 4 图3对应的初至时窗聚类结果

Figure 4. Clustering result of time windows corresponding to Fig.3

下载: 全尺寸图片幻灯片

图 5 理论模型经长短时窗能量比法和加权K均值方法的拾取结果

Figure 5. Picked results by STA/LTA （stars） and weighted K-means （solid circles） algorithms for theoretical records

下载: 全尺寸图片幻灯片

图 6 加权K均值（a）和长短时窗能量比（b）方法拾取直方图

Figure 6. Histogram of time error for theoretical records by weighted K-means （a） and STA/LTA （b） algorithms

下载: 全尺寸图片幻灯片

图 7 利用加权K均值（a）、BP神经网络（b）和STA/LTA （c）方法所得的实际资料初至拾取结果

Figure 7. First arrival time of a practical data based on weighted K-means （a），BP neural network （b） and STA/LTA （c） algorithms

下载: 全尺寸图片幻灯片

图 8 基于加权K均值（a）、BP神经网络（b）和STA/LTA （c）方法的拾取误差直方图

Figure 8. Histogram of time errors based on weighted K-means （a），BP neural network （b） and STA/LTA （c） algorithms

下载: 全尺寸图片幻灯片

图 9 实际微地震资料

Figure 9. Practical micro-seismic records

下载: 全尺寸图片幻灯片

图 10 第十道实际微地震资料（图中数字为三次震动事件）

Figure 10. The tenth trace of practical micro-seismic record where the figures with circle demonstrate three different micro-seismic events

下载: 全尺寸图片幻灯片

图 11 图10中三个震动事件的时窗局部放大图

黑色圆点为本文算法自动拾取的震动事件初至起跳时间

Figure 11. Zoom of the three micro-seismic events as shown in Fig.10 in time windows

Black dots represent first arrival’s jumping time picked by the algorithm used in this study

下载: 全尺寸图片幻灯片

图 12 震动事件拾取结果

Figure 12. Picking results of micro-seismic events

下载: 全尺寸图片幻灯片

表 1 三种算法性能分析

Table 1 Performances analyses of the three algorithms

方法	属性计算时间/s	算法运行时间/s	拾取准确率
加权K均值	35.65	0.168	100%
BP神经网络	35.65	5.127 6	64.91%
STA/LTA	−	4.394	35.68%

下载: 导出CSV

参考文献(30)

常旭,刘伊克. 2002. 地震记录的广义分维及其应用[J]. 地球物理学报,45(6):839–846. doi: 10.3321/j.issn:0001-5733.2002.06.011

Chang X,Liu Y K. 2002. The generalized fractal dimension of seismic records and its applications[J]. Chinese Journal of Geophysics,45(6):839–846 (in Chinese).

陈东,皮德常. 2009. 基于属性加权的改进K-means算法[J]. 电脑知识与技术,5(9):2412–2413.

Chen D,Pi D C. 2009. Improved K-means algorithm based on the attributes weighted[J]. Computer Knowledge and Technology,5(9):2412–2413 (in Chinese).

李辉峰,邹强,金文星. 2006. 基于边缘检测的初至波自动拾取方法[J]. 石油地球物理勘探,41(2):151–155.

Li H F,Zou Q,Jin W X. 2006. Method of automatic first breaks pick-up based on edge detection[J]. Oil Geophysical Prospec-ting,41(2):151–155 (in Chinese).

孟宇奇, 2018. 基于谱多流形聚类的微地震信号处理[D]. 长春: 吉林大学: 21−37.

Meng Y Q, 2018. Research on Microseismic Data Processing Based on Spectral Multi-Manifold Cluster Method[D]. Changchun: Jilin University: 21−37 （in Chinese）.

潘树林,高磊,周熙襄,钟本善. 2006. 基于单道边界检测和样条插值的初至波自动拾取[J]. 石油物探,45(3):245–249. doi: 10.3969/j.issn.1000-1441.2006.03.007

Pan S L,Gao L,Zhou X R,Zhong B S. 2006. Automatic method of first break picking based on edge detection and spline interpolation[J]. Geophysical Prospecting for Petroleum,45(3):245–249 (in Chinese).

裴正林,余钦范. 1999. 基于小波变换和BP神经网络的地震波初至拾取方法[J]. 勘察科学技术,30(4):61–64. doi: 10.3969/j.issn.1001-3946.1999.04.015

Pei Z L,Yu Q F. 1999. A wavelet transform and BP neural network-based algorithm for detecting first arrivals on seismic waves[J]. Site Investigation Science and Technology,30(4):61–64 (in Chinese).

钱光萍. 2001. 基于分形维地震道初至拾取方法研究[D]. 成都: 成都理工大学: 17−42.

Qian G P. 2001. Study on Fractal-Based For Picking First Arrivals on Seismic Traces[D]. Chengdu: Chengdu University of Technology: 17−42 （in Chinese）.

王永刚, 乐友喜, 张军华. 2007. 地震属性分析技术[M]. 东营: 中国石油大学出版社: 32−82.

Wang Y G, Le Y X, Zhang J H. 2007. Seismic Attribute Analysis Technique[M]. Dongying: China University of Petroleum Press: 32−82 （in Chinese）.

魏超,郑晓东,李劲松. 2012. 基于量子蒙特卡罗的地震多属性聚类方法[J]. 石油地球物理勘探,47(5):747–753.

Wei C,Zheng X D,Li J S. 2012. The seismic multi-attribute clustering method based on quantum Monte Carlo method[J]. Oil Geophysical Prospecting,47(5):747–753 (in Chinese).

赵玄, 严家斌, 胡涛. 2018. 聚类分析在地球物理中的应用进展[J]. 中国科技信息, （15）: 103–105.

Zhao X, Yan J B, Hu T. 2018. Application of cluster analysis in geophysics[J]. China Science and Technology Information, （15）: 103–105 （in Chinese）.

朱丹. 2017. 基于FCM的微地震初至自动拾取算法研究[D]. 长春: 吉林大学: 9−33.

Zhu D. 2017. Study on Automatic Time Picking for Microseismic Data Based on FCM Algorithm[D]. Changchun: Jilin University: 9−33 （in Chinese）.

庄东海,肖春燕,颜永宁. 1994. 利用人工神经网络自动拾取地震记录初至[J]. 石油地球物理勘探,29(5):659–664.

Zhuang D H,Xiao C Y,Yan Y N. 1994. Seismic first arrival pickup using artificial neural network[J]. Oil Geophysical Prospec-ting,29(5):659–664 (in Chinese).

Chen Y K. 2018. Automatic microseismic event picking via unsupervised machine learning[J]. Geophys J Int,212(1):88–102. doi: 10.1093/gji/ggx420

Coppens F. 1985. First arrival picking on common-offset trace collections for automatic estimation of static corrections[J]. Geophys Prospect,33(8):1212–1231. doi: 10.1111/j.1365-2478.1985.tb01360.x

Harrington P. 2013. Machine Learning in Action[M]. Greenwich: Manning Publications: 184−198.

Maeda N. 1985. A method for reading and checking phase time in auto-processing system of seismic wave data[J]. Zisin,38(3):365–379. doi: 10.4294/zisin1948.38.3_365

McCormack M D,Zaucha D E,Dushek D W. 1993. First-break refraction event picking and seismic data trace editing using neural networks[J]. Geophysics,58(1):67–68. doi: 10.1190/1.1443352

Molyneux J B,Schmitt D R. 1999. First-break timing: Arrival onset times by direct correlation[J]. Geophysics,64(5):1492–1501. doi: 10.1190/1.1444653

施引文献(7)

期刊类型引用(2)

1.	张笑晗，张瑞青，朱敏. 四川盆地及邻区地壳分层结构研究——基于接收函数改进的序贯H-κ叠加方法的约束. 地球物理学报. 2024(08): 2946-2962 . 百度学术
2.	张瑞青，况春利，张笑晗，李永华. 沉积层结构被动源探测方法及其在典型盆地的应用. 地球与行星物理论评(中英文). 2023(01): 12-26 . 百度学术