测井技术2025，Vol.49Issue(1)：68-76,9.DOI:10.16489/j.issn.1004-1338.2025.01.008
基于电成像测井的碳酸盐岩地层地质结构自动分类新方法
Novel Automatic Classification Method for Geological Structures in Carbonate Formations Based on Electrical Imaging Logging
摘要
Abstract
In the processing of electrical imaging logging data for carbonate formations,it is challenging to distinguish mudstone laminae,natural fractures,induced fractures,and vugs due to their similar resistivities and overlapping occurrences.To address this issue,this paper proposes an automatic classification method based on skeleton maps and machine learning.First,an improved K-means clustering algorithm is used to segment regions of interest from the electrical imaging data.Then,image thinning and the Hough transform are employed to obtain a line-segment representation of the skeleton map,and line segments are grouped based on spatial geometric relationships.Next,a directional region-growing algorithm is applied to separate overlapping regions and to extract geometric features(e.g.,perimeter,area,average dip,etc.)of each independent connected region.Finally,these features are used to train an extreme gradient boosting(XGBoost)model,enabling the automatic classification of different geological structures.Validation using actual carbonate logging data shows that the classification accuracy of this method exceeds an 80%agreement rate with core observations.Compared with existing research,this study is the first to achieve automatic classification of mudstone laminae,natural fractures,induced fractures,and vugs in carbonate formations based on electrical imaging logs,significantly improving classification efficiency and accuracy.关键词
碳酸盐岩电成像/骨架图/方向性区域生长算法/极端梯度提升树/自动分类Key words
carbonate electrical imaging/skeletonization/directional region-growing algorithm/extreme gradient boosting/automatic classification分类
天文与地球科学引用本文复制引用
付雅峰,黄科,朱涵斌,王慧,张旭,赵洁,肖霓..基于电成像测井的碳酸盐岩地层地质结构自动分类新方法[J].测井技术,2025,49(1):68-76,9.基金项目
中国石油天然气集团有限公司科技项目"油基钻井液成像测井关键核心技术研究"(2024ZG45) （2024ZG45）
中国石油集团测井有限公司科学研究与技术开发项目"测井油藏与地质研究"(CNLC2023-8B03) （CNLC2023-8B03）
