基于邻近域特征的堆积层滑坡多维地电信息成像监测技术研究  

Multi-dimensional geoelectrical resistivity imaging monitoring for debris flow based on neighborhood domain features

作　　者：徐哈宁[1,2] 邓居智[1] 肖慧[2,3] XU Haning;DENG Juzhi;XIAO Hui(Engineering Research Center for Seismic Disaster Prevention and Engineering Geological Disaster Detection of Jiangxi Province,East China University of Technology,Nanchang 330013,China;Engineering Research Center of Nuclear Technology Application(Ministry of Education),East China Institute of Technology,Nanchang 330013,China;Jiangxi Engineering Laboratory on Radioactive Geoscience and Big Data Technology,East China University of Technology,Nanchang 330013,China)

机构地区：[1]东华理工大学江西省防震减灾与工程地质灾害探测工程研究中心,江西南昌330013 [2]东华理工大学核技术应用教育部工程研究中心,江西南昌330013 [3]东华理工大学江西省放射性地学大数据技术工程实验室,江西南昌330013

出　　处：《地学前缘》2023年第6期473-484,共12页Earth Science Frontiers

基　　金：江西省防震减灾与工程地质灾害探测工程研究中心开放基金项目(SDGD202005);核技术应用教育部工程研究中心基金项目(HJSJYB2019-7,HJSJYB2018-3);江西省自然科学基金项目(20212BAB203004);江西省放射性地学大数据技术工程实验室基金项目(JELRGBDT202206)。

摘　　要：堆积层滑坡作为一种非线性系统变形演化规律复杂,准确量化坡体形变过程中滑移面边界变化、降雨入渗边界、水头边界和流量边界等重要参数是堆积层滑坡变形破坏机理研究的基础。此项研究基于电阻率成像技术,可依据岩土体成分、结构的不同和地层间电性的差异实现快速、多维剖面成像,并通过深度神经网络建立了浅层高分辨率电阻率成像数据和各类监测数据的“时间-空间-属性”邻近域特征信息统一表达矩阵,实现了各监测项空间权重的快速收敛。在此基础上,用逐层、逐个递推式回归分析方法分析深层电阻率成像数据,准确构建了反映堆积层滑坡形变过程的内部结构多维电性特征模型。试验证明,该方法有效地提高了各类边界参数基于地电信息成像监测结果的实时性和准确性,为堆积层滑坡形变破坏机理研究提供理论与方法支撑。As a nonlinear system debris flows exhibit complex slope deformation patterns,and accurate quantification of the key parameters such as slip surface boundary change,rainfall infiltration boundary,head boundary and flow boundary during slope deformation is essential to studying the deformation mechanism of debris flows.Electrical resistivity imaging allows fast and multidimensional profile imaging based on compositional/structural differences between geotechnical bodies and difference in electrical property between strata.In this study,a dataset containing the“time-space-attribute”neighborhood parameters extracted from high-resolution shallow surface resistivity imaging data and various monitoring data was constructed by deep neural network learning,and used as input to generate,with rapid convergence,the relevant spatial weights matrix for monitoring items.On this basis,the deep resistivity imaging data were analyzed recursively,layer by layer,and a multidimensional internal structural model of the studied debris flow was accurately constructed based on the electrical characteristics of the debris flow to quantify its slope deformation process.This method was validated experimentally,which showed effective improvements in the real-timeliness and accuracy of various boundary parameters derived from geoelectrical resistivity imaging monitoring data.Results from this study have theoretical and methodological significances for understanding the slope deformation mechanism of debris flows.

关 键 词：堆积层滑坡 电阻率成像技术 邻近域特征 深度神经网络 

分 类 号：P642.22[天文地球—工程地质学] TP391[天文地球—地质矿产勘探] TP183[天文地球—地质学]