基于HDBSCAN算法的岩体结构面产状识别及分组方法

精准获取岩体结构面的产状信息是进行岩体稳定性分析工作的首要和关键步骤.针对基本DBSCAN算法在处理点云数据时存在聚类判据不足、参数敏感性较强等问题，提出了基于点云附加属性扩展聚类判据的HDBSCAN算法，旨在提高算法聚类的准确性及鲁棒性.同时，基于点云的颜色及密度属性实现了对自然状态下非全裸露岩体研究区域的分割和提取，减少非岩体结构面点云对结构面产状识别和分组的影响.将该方法应用于大石洞灰岩矿某一典型覆土岩坡，结果表明:该方法能够有效地剔除非目标点云数据，同时结构面产状提取和优势分组结果令人满意.与人工测量结构面产状方法相比，最大相对误差小于0.59%，具有一定的工程实用价值.

Identification and Grouping Method of Strike Information of Rock Mass Based on the HDBSCAN Algorithm

Accurate acquisition of the strike information of the rock discontinuities is the first and key step in rock stability analysis. To solve the problems of insufficient clustering criteria and strong parameter sensitivity of the basic DBSCAN algorithm in processing the point cloud data， the HDBSCAN algorithm was proposed to extend the clustering criteria based on additional attributes of point clouds to improve the accuracy and robustness of the clustering algorithm. Meanwhile， based on the color and density attributes of the point cloud， the segmentation and extraction of the studied area of non-fully exposed rock in natural state were realized and the influence of the point cloud at non-rock structural planes on the identification and grouping of rock discontinuities production was reduced. The method was applied to a typical overburden rock slope of Dashidong tuff mine， and the results show that the method proposed can effectively eliminate non-target point cloud data， while the maximum relative error is less than 0.59% compared with the manual method， the method of rock discontinuities strike extraction and grouping are satisfactory and has practical engineering value.