基于均值漂移聚类算法的岩体结构面产状优势分组研究
DOI:
作者:
作者单位:

1.重庆交通大学河海学院;2.山西省勘察设计研究院有限公司

作者简介:

通讯作者:

中图分类号:

TD679

基金项目:

重庆市自然科学基金面上项目(cstc2021jcyj-msxmX0869);重庆市教委科学技术研究项目(KJQN202200709);重庆市研究生联合培养基地建设项目(JDLHPYJD2022004)


Study on Predominance Grouping of Rock Mass Structural Plane Occurrence Based on Mean Shift Clustering Algorithm
Author:
Affiliation:

1.Hohai College, Chongqing Jiaotong University;2.Shanxi Survey and Design Institute Co

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    岩体结构面产状的优势分组对于揭示不同类型结构面的分布规律和特征具有重要意义。传统的结构面极点密度图分组方法通常较为依赖地质经验,缺乏一定客观性,为此,本文引入均值漂移聚类算法开展岩体结构面产状优势分组研究。首先,人工生成了不同离散程度岩体结构面产状数据。随后,将生成的产状数据转换为三维空间中的坐标,并以单位法向量的夹角正弦值作为相似性度量标准。接下来采用均值漂移算法对度量的数据集进行聚类分析,通过与传统的极点密度图法和K均值聚类算法进行比较,有效性检验指标和聚类错误识别率与K均值聚类算法接近一致。最后以重庆三功矿岩质边坡为工程实例,通过野外采集到的结构面数据验证了新方法的合理性及有效性。研究结果表明:该方法聚类效果优于传统的极点图分组方法和K均值聚类算法,聚类结果客观合理,对近水平产状也有良好的聚类效果。

    Abstract:

    The dominant grouping of rock mass structural plane occurrence is of great significance to reveal the distribution and characteristics of different types of structural plane. The traditional grouping method of structural plane pole density map usually relies on geological experience and lacks some objectivity. Therefore, the mean shift clustering algorithm is introduced in this paper to study the predominance grouping of rock mass structural plane occurrence. Firstly, the occurrence data of rock mass structural plane with different degrees of dispersion are generated manually. Then, the generated occurrence data is converted into coordinates in three-dimensional space, and the sinusoidal value γ of the unit normal vector is used as the similarity measure. Next, the mean shift algorithm is used to perform cluster analysis on the measured data set. Compared with the traditional pole density map method and K-means clustering algorithm, the validity test index and clustering error recognition rate are close to the K-means clustering algorithm. Finally, taking the Chongqing Sangong ore-rock slope as an engineering example, the rationality and effectiveness of the new method are verified by the field data collected from the structural plane. The results show that the clustering effect of the proposed method is superior to the traditional pole map grouping method and K-means clustering algorithm. The clustering results are objective and reasonable, and the clustering effect for near-horizontal occurrence is also good.

    参考文献
    相似文献
    引证文献
引用本文

彭是焱,周鑫,申壮,等. 基于均值漂移聚类算法的岩体结构面产状优势分组研究[J]. 科学技术与工程, , ():

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-12-06
  • 最后修改日期:2024-06-14
  • 录用日期:2024-06-24
  • 在线发布日期:
  • 出版日期:
×
亟待确认版面费归属稿件,敬请作者关注
《科学技术与工程》入选维普《中文科技期刊数据库》自然科学类期刊月度下载排行榜TOP10