基于模糊C均值算法改进的岩体结构面识别与信息提取  

Rock mass structural plane identification and feature extraction with improved fuzzy C-means algorithm

作　　者：李斌斌 王新辉 王学良[3,4] 金超 孙娟娟 王赞 LI Binbin;WANG Xinhui;WANG Xueliang;JIN Chao;SUN Juanjuan;WANG Zan(School of Earth Sciences and Engineering,Hebei University of Engineering,Handan,Hebei 056038,China;Beijing North-Star Technology Development Co.,Ltd.,Beijing 100029,China;Key Laboratory of Shale Gas and Geoengineering,Institute of Geology and Geophysics,Chinese Academy of Sciences,Beijing 100029,China;College of Earth and Planetary Sciences,University of Chinese Academy of Sciences,Beijing 100049,China;State Key Laboratory of Precision Blasting,Jianghan University,Wuhan,Hubei 430056,China;Hubei Key Laboratory of Blasting Engineering,Jianghan University,Wuhan,Hubei 430056,China)

机构地区：[1]河北工程大学地球科学与工程学院,河北邯郸056038 [2]北京洛斯达科技发展有限公司,北京100029 [3]中国科学院地质与地球物理研究所页岩气与地质工程重点实验室,北京100029 [4]中国科学院大学地球与行星科学学院,北京100049 [5]江汉大学精细爆破国家重点实验室,湖北武汉430056 [6]江汉大学爆破工程湖北省重点实验室,湖北武汉430056

出　　处：《公路交通科技》2025年第3期126-135,共10页Journal of Highway and Transportation Research and Development

基　　金：国家自然科学基金项目(42172304)。

摘　　要：【目标】传统的模糊C均值算法在岩体结构面识别过程中存在明显的缺陷,即对出现的噪声和离群点较为敏感。此外,当前在岩体结构面间距和迹长等几何特征提取过程中,通常依赖手动方法或基于点云的技术。这些方法容易受到外部因素的干扰,从而影响结果的准确性。为解决上述问题,提出了一套岩体结构面信息自动提取的方法。【方法】首先,提出一种改进的模糊C均值算法。该算法通过引入新的参数来自适应地调节算法的模糊度,从而增强算法的鲁棒性并优化聚类效果。在此基础上,利用Fast Marching算法识别出闭合结构面,并结合Matlab与ArcGIS软件,构建结构面几何特征参数的自动化提取方法,实现对结构面产状、间距和迹长的自动测量与统计。【结果】实际岩体结构面产状的标准差最高减小了4°,结构面间距与迹长的最大误差分别为1.1 cm和3.4 cm,满足区域信息提取的误差容许要求。【结论】相较于传统方法,所提出的算法在结构面信息的提取过程中展现出较高的精度,且大幅提高了效率,能够更好地应用于实际工程中,为岩体结构面分析提供了一种高效、准确的技术方案。[Objective]The traditional fuzzy C-means algorithm exhibits significant deficiencies when identifying rock mass structural planes,particularly due to its high sensitivity to noise and outliers.Furthermore,the current methods for extracting geometric features,e.g.,rock structural plane spacing and trace lengths,often rely on manual techniques or point cloud-based technologies.These methods can be adversely affected by external factors,thereby compromising the accuracy.To solve these mentioned problems,a novel approach for automatic extraction of rock structural plane information was proposed.[Method]First,an improved fuzzy C-means algorithm was proposed.This algorithm adaptively adjusted the algorithm ambiguity by introducing new parameters to enhance the algorithm robustness,and optimize the clustering effect.On this basis,the closed structural planes were identified by employing Fast Marching algorithm.Combining with Matlab and ArcGIS software,an automatic extraction method for structural plane geometric feature parameters was proposed to realize the automatic measurement and statistics of structural plane occurrence,spacing,and trace length.[Result]The standard deviation of actual rock mass structural plane occurrence is reduced by 4°at most.The maximum errors of structural plane spacing and trace length are 1.1 cm and 3.4 cm respectively,which meet the error tolerance requirements of regional information extraction.[Conclusion]Compared with traditional methods,the proposed algorithm exhibits the superior accuracy in extracting structural plane information and significantly enhances efficiency,thus rendering it more suitable for application in real engineering projects.Additionally,it offers an effective and precise technical solution for analyzing rock structural planes.

关 键 词：道路工程 自动提取 模糊C均值算法 结构面识别 几何特征 岩体结构 

分 类 号：U412.22[交通运输工程—道路与铁道工程]