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

doi:10.13745/j.esf.sf.2023.2.60

地学前缘 ›› 2023, Vol. 30 ›› Issue (6): 473-484.DOI: 10.13745/j.esf.sf.2023.2.60

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

徐哈宁¹^,²(), 邓居智¹, 肖慧²^,³^,^*()

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

收稿日期:2022-08-10 修回日期:2023-03-01 出版日期:2023-11-25 发布日期:2023-11-25
通讯作者: * 肖慧(1978—),女,副教授,硕士生导师,主要从事光栅光纤在地质灾害监测领域的应用研究。E-mail: xhandxhn@163.com
作者简介:徐哈宁(1979—),男,副教授,硕士生导师,主要从事地质灾害监测技术研究。E-mail: 33007063@qq.com
基金资助:
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);核技术应用教育部工程研究中心基金项目(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)

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

XU Haning¹^,²(), DENG Juzhi¹, XIAO Hui²^,³^,^*()

1. Engineering Research Center for Seismic Disaster Prevention and Engineering Geological Disaster Detection of Jiangxi Province, East China University of Technology, Nanchang 330013, China
2. Engineering Research Center of Nuclear Technology Application (Ministry of Education), East China Institute of Technology, Nanchang 330013, China
3. Jiangxi Engineering Laboratory on Radioactive Geoscience and Big Data Technology, East China University of Technology, Nanchang 330013, China

Received:2022-08-10 Revised:2023-03-01 Online:2023-11-25 Published:2023-11-25

摘要/Abstract

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。

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

Abstract:

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.

Key words: debris-flow, resistivity imaging technique, neighborhood domain feature, deep neural network

中图分类号:

徐哈宁, 邓居智, 肖慧. 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[J]. 地学前缘, 2023, 30(6): 473-484.

XU Haning, DENG Juzhi, XIAO Hui. Multi-dimensional geoelectrical resistivity imaging monitoring for debris flow based on neighborhood domain features[J]. Earth Science Frontiers, 2023, 30(6): 473-484.

图/表 16

图1 黄石市阳新县木港镇石溪村卫生所屋后滑坡现场照片(左)和监测点布设图(右)

Fig.1 Topographic features of the debris-flow deposit in Huangshi City (left), and planar layout of monitoring stations (right)

图2 降雨量和近地表含水率监测结果

Fig.2 Bi-hourly rainfall (blue) and near-surface water content (orange) by GNSS systems

图3 JC1监测点数据

Fig.3 GNSS monitoring data by JC1, showing displacement time series in the northward (blue), eastward (red) and downward (yellow) directions

图4 JC2监测点数据

Fig.4 GNSS monitoring data by JC2

图5 滑坡现场俯视图

Fig.5 Top view of the debris-flow deposit, showing placement of GNSS systems

图6 电阻率数据采集样本点的二维空间邻近性关系

Fig.6 Two-dimensional spatial proximity relationships among resistivity measuring points

图7 邻近域深度神经网络非线性回归分析实现过程

Fig.7 Methodology of nonlinear regression analysis with neighborhood deep neural network

图8 增设的电阻率成像监测系统现场布设图

Fig.8 Position of the added resistivity imaging monitoring system in relation to GNSS systems

图9 降雨量和土壤含水率数据图

Fig.9 Daily rainfall and soil water content

图10 3月20日电阻率成像监测反演结果

Fig.10 2D inversion of the March 20th resistivity imaging monitoring data

图11 深度神经网络对数据集的误差分析

Fig.11 Error analysis for training dataset

图12 观测数据与三种模型分析结果比对

Fig.12 Comparison of monitoring data with three regression models

图13 三种算法的误差对比图

Fig.13 Comparison of error rates for three algorithms

表1 主要指标参数

Table 1 Performance evaluation of three regression models

分析方法	RMSE	MAE	MSE	R²
线性回归模型	0.847	0.887	0.807	0.913
非线性回归模型	0.550	0.405	0.550	0.959
邻近域深度神经网络非线性回归模型	0.265	0.283	0.298	0.992

图14 两个数据集回归分析比对

Fig.14 Comparison of model accuracy between two input datasets

图15 两种数据集回归分析结果和观测结果对比 a—数据集2回归分析成像结果;b—数据集1回归分析成像结果;c—观测值成像结果。

Fig.15 Comparison of resistivity tomography models with imaging data between two input datasets

参考文献 25

[1]	贺可强, 周敦云, 王思敬. 降雨型堆积层滑坡的加卸载响应比特征及其预测作用与意义[J]. 岩石力学与工程学报, 2004, 23(16): 2665-2670.
[2]	钟源, 李远耀, 殷坤龙, 等. 基于物理模型试验的厚层堆积层滑坡强降雨触发机制[J/OL]. 地球科学: 1-14[2022-02-14]. https://kns-cnki-net.webvpn.ecut.edu.cn/kcms/detail/42.1874.P.20220113.1040.002.html.
[3]	李高, 谭建民, 王世梅, 等. 滑坡对降雨响应的多指标监测及综合预警探析: 以赣南罗坳滑坡为例[J]. 地学前缘, 2021, 28(6): 283-294. DOI
[4]	赵超英, 刘晓杰, 张勤, 等. 甘肃黑方台黄土滑坡InSAR识别、监测与失稳模式研究[J]. 武汉大学学报(信息科学版), 2019, 44(7): 996-1007.
[5]	FERMIN V, JOSE T, JOSE A R, et al. New framework based on fusion information from multiple landslide data sources and 3D visualization[J]. Journal of Earth Science, 2020, 31(1): 159-170. DOI
[6]	张明, 胡瑞林, 殷跃平, 等. 川东缓倾红层中降雨诱发型滑坡机制研究[J]. 岩石力学与工程学报, 2014, 33 (增刊2): 3783-3790.
[7]	LAPENNA V, LORENZO P, PERRONE A, et al. 2D Electrical resistivity imaging of some complex landslides in Lucanian Apennine Chain, southern Italy[J]. Geophysics, 2005, 70(1): 11-18.
[8]	孙永帅, 胡瑞林. 不同角度基覆面上土石混合体变形试验研究及对滑坡演化的启示[J]. 地学前缘, 2023, 30(3): 494-504. DOI
[9]	陈善雄, 许锡昌, 徐海滨. 降雨型堆积层滑坡特征及稳定性分析[J]. 岩土力学, 2005, 26(增刊2): 6-10.
[10]	郭璐, 贺可强, 贾玉跃. 水库型堆积层滑坡位移方向协调性参数及其失稳判据研究[J]. 水利学报, 2018, 49(12): 1532-1540.
[11]	何少其, 刘元雪, 杨骏堂, 等. 库岸堆积层滑坡位移的分量响应模式及多因子模型[J]. 岩土力学, 2020, 41(8): 2773-2784.
[12]	杨天春, 陈卓超, 梁竞, 等. 天然电场选频测深法在地下水勘探中的异常理论分析与实践应用[J]. 地学前缘, 2020, 27(4): 302-310. DOI
[13]	闫亚景, 闫永帅, 赵贵章, 等. 基于高密度电法的天然边坡水分运移规律研究[J]. 岩土力学, 2019, 40(7): 2807-2814.
[14]	SUNARYO A S, ALAMSYAH M, YUWONO, et al. Slope stability analysis for landslides natural disaster mitigation by means of geoelectrical resistivity data in Gedangan of South Malang[J]. Materials Science and Engineering, 2019, 546(2): 334-346.
[15]	AKMANM H, AMIR A, PUTRA, et al. Implementation of least-square constrain inversion method of geoelectrical resistivity data Wenner-Schlumberger for investigation the characteristic of landslide[J]. Indian Journal of Public Health Research & Development, 2019, 10(8): 2456-2469.
[16]	GURUH S, MUHAMMAD F R, JUNUN S, et al. Characterization of displacement and internal structure of landslides from multitemporal UAV and ERT imaging[J]. Landslides, 2020, 17(10): 2455-2468. DOI
[17]	罗鸿东, 李瑞冬, 张勃, 等. 基于信息量法的地质灾害气象风险预警模型: 以甘肃省陇南地区为例[J]. 地学前缘, 2019, 26(6): 289-297. DOI
[18]	ZHANG L, SHI B, ZHU H, et al. PSO-SVM-based deep displacement prediction of Majiagou landslide considering the deformation hysteresis effect[J]. Landslides, 2021, 18(1): 179-193. DOI
[19]	CAITLIN J. TRAN, OMAR E, et al. Unsupervised classification for landslide detection from airborne laser scanning[J]. Geosciences, 2019, 9(5): 113-124. DOI URL
[20]	MEENA S R, GHORBANZADEH O, WESTEN C, et al. Rapid mapping of landslides in the Western Ghats (India) triggered by 2018 extreme monsoon rainfall using a deep learning approach[J]. Landslides, 2021, 18(5): 1937-1950. DOI
[21]	谢顺平, 叶罕霖. 面向空间关系复合的矢量多边形图形拓扑叠置分析算法[J]. 计算机辅助设计与图形学学报, 2018, 30(9): 1678-1685.
[22]	王中一. 顾及全域时空复杂邻近性的神经网络加权回归方法[D]. 杭州: 浙江大学, 2020.
[23]	WU B, LI R, HUANG B. A geographically and temporally weighted auto regressive model with application to housing prices[J]. International Journal of Geographical Information Science, 2014, 28(5): 1186-1204. DOI URL
[24]	许泽坤, 陈隽. 非线性结构地震响应的神经网络算法[J]. 工程力学, 2021, 38(9): 133-145.
[25]	艾骁. 基于机器学习的地震滑坡易发性评估模型构建[D]. 哈尔滨: 中国地震局工程力学研究所, 2021.

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

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 16

参考文献 25

相关文章 9

编辑推荐

Metrics

本文评价

[1]	李高, 谭建民, 王世梅, 林旭, 陈勇, 王力, 郭飞. 滑坡对降雨响应的多指标监测及综合预警探析:以赣南罗坳滑坡为例[J]. 地学前缘, 2021, 28(6): 283-294.
[2]	陈剑, 陈瑞琛, 崔之久. 高速远程滑坡的地貌学与沉积学研究进展[J]. 地学前缘, 2021, 28(4): 349-360.
[3]	张永双, 刘筱怡, 吴瑞安, 郭长宝, 任三绍. 青藏高原东缘深切河谷区古滑坡:判识、特征、时代与演化[J]. 地学前缘, 2021, 28(2): 94-105.
[4]	殷志强, 魏刚, 秦小光, 李文娟, 赵无忌. 青藏高原东北缘黄河上游滑坡与堰塞湖研究进展[J]. 地学前缘, 2021, 28(2): 46-57.
[5]	孙启良, 解习农, 吴时国. 南海北部海底滑坡的特征、灾害评估和研究展望[J]. 地学前缘, 2021, 28(2): 258-270.
[6]	白世彪, 崔鹏, 葛永刚, 王昊. 古滑坡测年方法与定年精度的提高途径[J]. 地学前缘, 2021, 28(2): 19-34.
[7]	陈晓利, 单新建, 张凌, 刘春国, 韩娜娜, 兰剑. 地震诱发滑坡的快速评估方法研究:以2017年M_S7.0级九寨沟地震为例[J]. 地学前缘, 2019, 26(2): 312-320.
[8]	王磊, 李滨, 高杨, 朱赛楠. 大型厚层崩滑体运动特征模拟研究：以重庆武隆县羊角场镇大巷危岩为例[J]. 地学前缘, 2016, 23(2): 251-259.
[9]	袁仁茂, 谭锡斌, 陈桂华, 徐锡伟. 地震破裂带特殊部位大型滑坡及其基于构造地貌发生模型的机制解释:以东河口抛射型滑坡为例[J]. 地学前缘, 2010, 17(5): 243-253.