电力勘测设计Issue(6)：48-52,5.DOI:10.13500/j.dlkcsj.issn1671-9913.2025.06.010
基于多重聚类算法的岩体结构面快速识别研究
Research on Rapid Identification Based on Multiple Clustering Algorithm for Structural Plane of Rock Mass
摘要
Abstract
The traditional methods for identifying structural planes of rock mass have problems such as reliance on manual experience and low computational efficiency.To solve the problems,an exploration adit in a pumped storage power station in northwestern China is chosen as a study case.A rapid automatic identification method based on multiple clustering algorithms proposed,with the purpose of enhancing both the objectivity and efficiency of structural plane recognition.3D laser point cloud data are used to create combined feature vectors.Both K-means and DBSCAN clustering methods are applied to automatically identify rock mass structural planes.An empirical threshold is set to filter out fractured surfaces.The research results show that this method can efficiently and automatically accomplish the structural plane identification and the statistical analysis of attitude information for rock mass,significantly reducing manual intervention.The calculation speed is better than the traditional method and can meet the practical needs of engineering.关键词
抽水蓄能电站/多重聚类/三维激光点云/岩体结构面/特征向量Key words
pumped storage power station/3D laser point cloud/rock mass structural plane/feature vector/multiple clustering分类
能源科技引用本文复制引用
任晖,张海龙,郭亚真,库新勃,李龙威,贺鸣..基于多重聚类算法的岩体结构面快速识别研究[J].电力勘测设计,2025,(6):48-52,5.基金项目
中国能源建设股份有限公司科技项目"北斗在能源电力领域的应用技术研究"(CEEC2022-ZDYF-01) （CEEC2022-ZDYF-01）
中国电力工程顾问集团有限公司集中开发科研项目"基于人工智能的电力工程勘测数据辅助应用关键技术研究"(DG2-L03-2024) （DG2-L03-2024）
中国能源建设股份有限公司重点研发项目"无人机激光雷达技术在电力工程测绘中的应用研究"(NW-RD036-2023). （NW-RD036-2023）
