| Most of the disposal of radioactive nuclear waste adopts deep burial,so it has great significance to make further research on the development characteristics of rock mass discontinuities in high level radioactive waste disposal sites.In this paper,a national nuclear waste disposal laboratory in northwestern China is taken as the research object,the 3D digital surface models of surface and underground tunnel rock mass in research site are reconstructed based on the digital photogrammetry technology and related computer vision algorithm,and the digital extraction of discontinuities parameters and 3D fractured network simulation are realized.The research work and corresponding conclusions are summarized as follows:(1)The 3D digital surface models of rock mass in the study area are reconstructed using digital photogrammetry technology and SFM 3D modeling algorithm.The step of rock mass DSM reconstruction is first collecting rock mass images,then matching image features based on SIFT algorithm,reconstructing sparse point cloud,encrypting point cloud,and reconstructing the rock mass surface model using MeshLab which is a mesh editing software.The reconstructed DSM has high accuracy,and the errors of control points are all less than1 cm,which can ensure the accuracy of subsequent discontinuity parameters extraction results.(2)A method is presented to automatic recognize discontinuity based on 3D digital surface model of rock mass in this paper.The main flow of the discontinuity recognition method is smoothing the digital surface model to reduce the noise on the surface of the rock mass model by the Laplacian smoothing algorithm,establishing the flatness index to segment the model by the principal component analysis,finally searching the discontinuity based on the regional growth principle to obtain the independent and complete discontinuity planes.The proposed method was subsequently applied to detect and identify discontinuities of a typical exposed rock slope and a large-scale underground rock tunnel with large-scale.Compared with the manual field measurement and the mainstream discontinuity identification software(Discontinuity Set Extractor,DSE),the result shows that the proposed method not only has good applicability of the rock mass discontinuity exposed on the ground,but also performs well on the underground tunnel discontinuity identification.(3)The random sampling consensus(RANSAC)algorithm is used to fit the point cloud of the discontinuity plane and compared with the PCA method,the result shows that the RANSAC algorithm considers the fluctuation of the points and is more suitable for the fitting of the discontinuity plane point cloud.An improved fuzzy clustering algorithm program is coded based on maximum density principle.The 3D trace models of outcrops and underground tunnel are established based on the measured data in the study area,and the parameters of discontinuity like trace length,spacing and intensity are extracted digitally by coding corresponding programs.(4)The 3D network simulation of discontinuity is realized based on the digitally extracted discontinuity parameters of rock mass in the study area.The discontinuity shape in the 3D network model is disk,the diameter of disks conforms to lognormal distribution,and the center location of disks conforms to poisson uniform distribution.Based on differential theory,the problem of intersection between disk and tunnel surface is transformed into the problem of intersection between straight line and disk,which greatly improves the computational efficiency of the program and successfully realizes the visualization of simulated trace on the tunnel model surface.In this paper,the application range of statistical discontinuity parameters of the circular window arranged on tunnel surface is analyzed.Based on the known discontinuity parameters obtained from the underground tunnel surface statistics,the discontinuity diameter and volume density are modified to make the final simulation 3D network model more consistent with the reality. |
PDF Full Text Request