Research On The Photogrammetric Interpretation Of Rock Mass Structural Plane Parameters And The Development And Application Of 3D Network Modeling System

In this thesis,the rock mass structure of the high-level radioactive waste disposal area in Beishan,Gansu Province is mainly studied.With the manual contacted measurement,a large number of structural plane information is obtained by means of photogrammetry.Based on this,the main research and conclusions are as follows:(1)Complete the digital acquisition of discontinuities parameters and statistical analysis of geometric parameter distribution.A large number of outcrop images of rock masses were obtained by photogrammetry.Image feature matching was performed using SIFT algorithm to obtain two-dimensional coordinates of fractures.Three-dimensional coordinates of the traces can be output by importing three-dimensional reconstruction software.On this basis,the calculation and grouping of discontinuity,the calculation and statistical analysis of the lengths of discontinuity,and the estimation of areal densities are completed.(2)A 3D stochastic model of the discontinuity was established and the DE algorithm was used to optimize the parameters of the radius and density of the structural surface.Monte Carlo method was used to randomly generate various parameters required for modeling,and a 3D random network model of discontinuity was initially established.The necessity of optimizing model parameters was compared between the measured trace length and the model cross-track length.The differential structure was used to optimize the structural surface radius and discretized bulk density.The accuracy has increased about 50%by comparing the cross-cutting traces and the measured traces of the optimized model,and the optimization effect is obvious.(3)The deterministic information is added to the pure stochastic model to establish a more realistic random-deterministic coupling model.Based on the actual measured fracture traces,deterministic discontinuity information is obtained and a random model is inserted according to the "nearest neighbor principle" to obtain a random-determined coupling model.Improved CS algorithm for model validation to verify the effectiveness of the coupled model.(4)Based on the stochastic-deterministic coupling model,the research and analysis of the excavation block excavation in underground roadway are completed.A model of underground excavation roadway is established.All discontinuity intersecting the roadway in the coupling model are searched for further elimination of redundant discontinuity,then search for closed loops and search for closed blocks.During the searching process of closed blocks,while the blocks overlap and contain,the blocks are screened and removed according to the principle of maximum volume,and the required blocks are finally obtained.The probability of blocks appearing on the surface of the roadway is calculated to provide a reference for the stability evaluation of the country rock.(5)Development software for obtaining geometric parameters of discontinuity and generation and optimization of 3D network model.Based on the VS2010 platform,all code is written in the C++ language under the MFC framework,and related visualizations are realized using OpenGL.The software can completely realize the acquisition of image feature-matching trace coordinates,calculation of fracture origin,calculation of grouping and areal density,and establishment of discontinuity 3D network model,adding parameters optimization module and block search applications to provide effective reference for the digitization of structural surfaces.