基于激光测量和FKM聚类算法的隧洞岩体结构面的模糊群聚分析

岩体隧洞结构面的数量、产状和发育情况是评价隧洞透水性和稳定性的重要指标。采用三维激光测量技术可以快速获取隧洞周围岩体激光点云数据和结构面的几何信息。提出了基于柱面投影和Delaunay生长算法的激光点云的三角网重构技术，利用构建的三角网平面方程的法向向量计算出单个三角网平面的产状，并应用FKM聚类算法进行隧洞岩体结构面产状的模糊群聚分类。该方法可以快速得到不同组的优势产状，且分类结果可用不同颜色进行空间显示。最后运用自主研发的FSS软件对某隧洞的岩体结构面优势产状进行了研究，研究结果表明：1）基于柱面投影和Delaunay生长法的点云构网方法特别适合于隧洞工程，构网速度快、效率高，重构得到的三角网表面与被采样的岩体表面拓扑差别最小；2）FKM模糊群聚方法对隧洞岩体结构面的群聚分类结果可靠、合理，具有较强的通用性和推广性。

Fuzzy Cluster Analysis of Rock Mass Discontinuity of Tunnel Based on Laser Measurement and FKM Clustering Algorithm

The number, attitude and developmental state of rock mass discontinuity of tunnel are important indexes to evaluate permeability and stability of tunnel. Laser point cloud data and geometry information of rock mass discontinuity around the tunnel can be obtained quickly using the three dimensional measurement technology. The reconstruction technology of triangular mesh of laser point cloud was proposed based on the projection of cylinder method and Delaunay growth algorithm, the attitude of single triangular mesh surface was calculated using the normal vector of the plane equation of the constructed triangular mesh, and fuzzy cluster classification of the attitude of the discontinuity of tunnel was carried out with the application of FKM (Fuzzy K-Mean) cluster method. The different groups of the attitude in advantage can obtained quickly using this method, and the spatial display of the result of the classification can be conducted with different colors. Finally, the study of the attitude in advantage of the discontinuity of some tunnel was carried out using the software FSS developed independently, the results of the study show that: 1)The constructing mesh with point cloud method based on the projection of cylinder method and Delaunay growth algorithm is particularly suitable for tunnel project, the speed of construction of mesh is fast, the efficiency is high, and the topology difference of the reconstructed surface of triangular mesh and the surface of sampling rock mass is minimum. 2)The result of cluster classification of the structure surface of rock mass of tunnel with the method of FKM fuzzy cluster is reliable, reasonable, and has strong commonality and extension.