岩质边坡节理信息数字识别及三维模型重构——以石家庄市南障城镇边坡为例  

Digital Recognition of Rock Slope Joint Information and 3D Model Reconstruction:Taking the Slope of Nanzhang Town in Shijiazhuang City as an Example

作　　者：刘宁 袁维 高岭[2] 白雪山 翟星 李建朋[2] 王安礼 徐江 LIU Ning;YUAN Wei;GAO Ling;BAI Xue-shan;ZHAI Xing;LI Jian-peng;WANG An-li;XU Jiang(School of Civil Engineering,Shijiazhuang Tiedao University,Shijiazhuang 050043,China;Hebei Provincial Communications Planning,Design and Research Institute Co.,Ltd.,Shijiazhuang 050300,China;Hebei Geo-Environment Monitoring,Shijiazhuang 050022,China;Hebei Key Laboratory of Geological Resources and Environment Monitoring and Protection,Shijiazhuang 050022,China;Engineering Technology Research Center for Water Resources Development and Utilization in Karst Regions of Guizhou Province,Guiyang 550002,China)

机构地区：[1]石家庄铁道大学土木工程学院,石家庄050043 [2]河北省交通规划设计院有限公司,石家庄050300 [3]河北省地质环境监测院,石家庄050022 [4]河北省地质资源环境监测与保护重点实验室,石家庄050022 [5]贵州省喀斯特地区水资源开发利用工程技术研究中心,贵阳550002

出　　处：《科学技术与工程》2023年第3期941-952,共12页Science Technology and Engineering

基　　金：河北省自然科学基金优秀青年基金(E2021210041);河北省重点研发计划(22375402D,22375420D);河北省教育厅重点项目(ZD2020333);贵州省科技支撑计划(黔科合支撑[2022]一般002)。

摘　　要：在地质调查中,为了实现岩质边坡节理信息数字识别及三维模型重构,开发设计了基于倾斜摄影的裂隙岩体边坡三维模型重构软件。该软件通过小型无人机倾斜摄影采集数据,然后基于增量式运动结构恢复算法和深度图融合的多视点立体视觉算法(multi-view stereo,MVS)进行三维模型重构,最后利用随机抽样一致算法(random sample consensus,RANSAC)和区域生长法对得到的结果进行岩体结构面识别,得到结构面产状等节理信息。把软件得到的三维模型和节理信息结合FLAC3D有限元软件,运用强度折减法进行稳定性分析。结果表明:采用小型无人机采集的数据完整有效;三维重构出的模型可以很好的描述出实际岩质边坡;结构面识别得到的岩质边坡节理信息误差较小;基于岩体体积节理数和结构面条件因子量化地质强度指标(geological strength index,GSI)的方法得到Hoek-Brown准则参数准确可靠。可见,利用小型无人机倾斜摄影结合开发设计的软件,可以快捷、高效和准确的实现岩质边坡节理信息数字识别及三维模型重构,对于实际工程具有一定的应用价值。In the geological survey,in order to realize the digital identification of joint information and 3D model reconstruction of rock slope,the 3D model reconstruction software of fractured rock slope based on tilt photography was developed and designed.The software collected data by tilt photography of a small unmanned aerial vehicle(UAV),and then reconstructed the three-dimensional model based on incremental motion structure recovery algorithm and depth map fusion multi-view stereo vision algorithm(multi-view stereo,MVS).Finally,the rock mass structural plane was identified by using random sampling consistent algorithm(random sample consensus,RANSAC)and region growing method,and the joint information such as the occurrence of structural plane was obtained.The three-dimensional model and joint information obtained by the software were combined with the FLAC3D finite element software,and the strength reduction method was used for stability analysis.The results show these as follows.the data collected by small UAV are complete and effective;the three-dimensional reconstruction model can well describe the actual rock slope;the error of joint information of rock slope obtained by structural plane identification is small;the Hoek-Brown criterion parameters are accurate and reliable based on the quantitative geological strength index(GSI)based on the number of rock mass volume joints and structural plane condition factors.It can be seen that the software developed by the combination of small UAV tilt photography and design can quickly,efficiently and accurately realize the digital identification of rock slope joint information and 3D model reconstruction,which has a certain application value for practical engineering.

关 键 词：边坡工程 摄影测量 三维模型重构 地质强度指标 强度折减法 

分 类 号：P231[天文地球—摄影测量与遥感]