基于PCA-SVM的砂砾岩岩性识别  

Lithologic Identification of Glutenite Based on PCA-SVM

作　　者：林香亮 朱建伟[1] 刘光寿 李世豪 陈雪娇 丁旭艳 袁瑞[1] 王德鹏 Lin Xiangliang;Zhu Jianwei;Liu Guangshou;Li Shihao;Chen Xuejiao;Ding Xuyan;Yuan Rui;Wang Depeng(School of Information and Mathematics,Yangtze University,Jingzhou 434023,Hubei;Tuha Branch of China National Petroleum Corporation Logging,Tuha 839000,Xinjiang)

机构地区：[1]长江大学信息与数学学院,湖北荆州434023 [2]中国石油集团测井有限公司吐哈分公司,新疆吐哈839000

出　　处：《长江大学学报（自然科学版）》2020年第1期21-26,I0002,I0003,共8页Journal of Yangtze University(Natural Science Edition)

基　　金：湖北教育厅科学研究计划项目“基于数字图像处理的沉积物粒度定量计算”(Q20181310);湖北省自然科学基金项目“过套管电阻率测井资料预处理方法研究及其软件开发”(2019CFB343);长江大学大学生创新创业训练计划项目“过套管电阻率测井异常值检验及其软件开发”(2018152)。

摘　　要：利用测井资料识别岩性实质是建立非线性数据模型的过程。表征复杂岩性的测井曲线相似度较高,对岩性识别存在一定的干扰,为此提出一种基于主成分分析(PCA)和支持向量机(SVM)的砂砾岩岩性识别预测模型:通过主成分分析对预处理后的测井数据进行主成分提取,利用提取的主成分建立支持向量机岩性识别预测模型。将测试样本数据输入模型中进行自动分类,最终该方法的岩性识别正确率达到87.07%。应用结果表明,将主成分分析与支持向量机相结合,在降低数据维度的同时,提高了岩性识别准确率,是一种实用且有效的识别复杂岩性的方法,具有一定的推广和应用价值。The essence of lithologic identification by logging data is a process of establishing nonlinear data model.The logging curves that characterize complex lithology have high similarity,which may interfere with the lithologic identification.A prediction model for lithologic identification of glutenite(PCA-SVM)based on principal component analysis(PCA)and support vector machine(SVM)was proposed.PCA was used to extract the principal components of the preprocessed logging data,and the prediction model for support vector machine lithologic identification by SVM was established by using the extracted principal components.The data of test samples were input into the model for automatic classification,and finally the accuracy of lithologic identification of this method reached 87.07%.The practical application results show that the combination of PCA with SVM can both reduce the data dimension and improve the accuracy of lithologic identification.It is a practical and effective method to identify the complex lithology,which has certain value of popularization and application.

关 键 词：砂砾岩 岩性识别 主成分分析(PCA) 支持向量机(SVM) 

分 类 号：P618.13[天文地球—矿床学]