摘要
地震数据的构造解释对矿山安全高效开采有重要作用,地震属性常用来进行构造解释,而单一属性及传统属性叠置方法无法完全利用地震数据中的信息。本文使用基于主成分分析-竞争神经网络的方法对多种地震属性进行融合聚类,实现复杂构造的识别。首先提取与构造相关性强的地震属性,然后利用主成分分析方法得到贡献率最大的几个主成分分量,最后利用无监督学习方法中的竞争神经网络来实现对选定的主成分分量的融合和聚类。以邢东矿区1200勘探区(经实际揭露为构造发育区域)的地震数据作为研究对象,应用基于主成分分析-竞争神经网络的多属性融合聚类方法进行分析,聚类图像能够清晰对应实际地质异常,有效分辨构造分布特征,为多属性构造识别提供了一种可行的方法。
The structural interpretation of seismic data plays an important role in the safe and efficient mining of mines.Seismic attributes are often used for structural interpretation,but the information in seismic data cannot be fully utilized by single attribute and traditional attribute superposition methods.In this paper,the method based on principal component analysis and competitive neural network was used to fuse and cluster various seismic attributes to realize the recognition of complex structures.Firstly,the seismic attributes with strong correlation with the structure were extracted,and then the principal component analysis method was used to obtain the principal component components with the largest contribution rate;finally,the competitive neural network in unsupervised learning method was used to realize the fusion and clustering of the selected principal component components.The seismic data of 1200 exploration area in Xingdong mining area was taken as the research object,the multi-attribute fusion clustering method based on principal component analysis and competitive neural network was applied.The final clustering image can clearly correspond to the actual geological anomalies and effectively distinguish the structural distribution characteristics,which provided a feasible method for multi-attribute structure recognition.
作者
姚江凯
刘家豪
Yao Jiangkai;Liu Jiahao(School of Geosciences&Surveying Engineering,China University of Mining and Technology,Beijing 100083,China)
出处
《煤炭与化工》
CAS
2020年第12期67-71,共5页
Coal and Chemical Industry
基金
国家重点研发计划项目(2018YFC0807800)。
关键词
无监督学习
地震属性
构造识别
属性融合
聚类分析
unsupervised learning
seismic attribute
structure recognition
attribute fusion
cluster analysis