基于卷积神经网络模型的区域滑坡敏感性评价——以川藏铁路沿线为例  

Evaluation of regional landslide susceptibility based on the convolution neural network model:a case study along the Sichuan Tibet Railway

作　　者：蒋万钰 陈冠 孟兴民 张毅 曾润强 岳东霞[2] 周自强[3] 赵岩 金佳成 梁懿文 JIANG Wan-yu;CHEN Guan;MENG Xing-min;ZHANG Yi;ZENG Run-qiang;YUE Dong-xia;ZHOU Zi-qiang;ZHAO Yan;JIN Jia-cheng;LIANG Yi-wen(Technology&Innovation Centre for Environmental Geology and Geohazards Prevention,School of Earth Sciences,Lanzhou University,Lanzhou 730000,China;Technology&Innovation Centre for Environmental Geology and Geohazards Prevention,College of Earth and Environmental Sciences,Lanzhou University,Lanzhou 730000,China;Geological Hazards Prevention Institute,Gansu Academy of Sciences,Lanzhou 730000,China)

机构地区：[1]兰州大学地质科学与矿产资源学院甘肃省环境地质与灾害防治技术创新中心,兰州730000 [2]兰州大学资源环境学院甘肃省环境地质与灾害防治技术创新中心,兰州730000 [3]甘肃省科学院地质自然灾害防治研究所,兰州730000

出　　处：《兰州大学学报（自然科学版）》2022年第2期203-211,共9页Journal of Lanzhou University(Natural Sciences)

基　　金：国家科技部重点研发课题项目(2017YFC1501005);第二次青藏高原综合科学考察研究项目(SQ2021QZKK0201);甘肃省科技重大专项计划项目(19ZD2FA002);甘肃省重点研发计划项目(20YF8FA074);甘肃省科技计划项目(18YF1WA114)。

摘　　要：在遥感解译、野外调查及前人研究的基础上,基于共线性诊断与极端随机数算法对13个影响因子的相关性和重要性进行评估,运用一维卷积神经网络构建川藏铁路沿线滑坡敏感性评价模型,使用受试者工作特征(ROC)曲线进行精度验证.结果表明,影响该区滑坡发育的主控因子为距河流距离、道路与断层的距离以及岩性,敏感性极低、低、中、高、极高的面积分别占总面积的66.30%、10.98%、7.55%、6.91%和8.26%.敏感性高的区域主要分布在河流、道路两侧和断层发育的区域;全线滑坡敏感性高低呈分段式分布,在山南-朗县段、波密-理塘段最为严重;一维卷积神经网络的ROC曲线下覆盖的面积为0.958,能有效地进行该区域的滑坡敏感性评价.Based on remote sensing interpretation,field surveys and previous studies,collinearity diagnosis and information gain was used to evaluate the correlation and importance of 13 impact factors,one dimensional convolutional neural network(1D CNN) employed to build an evaluation model,and receiver operating characteristic(ROC) curve for accuracy verification.The results showed that the main controlling factors affecting the development of landslides in this area were distance from rivers,roads,faults and lithological conditions.Areas with extremely low,low,medium,high and extremely high susceptibility accounted for 66.30%,10.98%,7.55%,6.91% and 8.26% of the total areas,respectively.The areas with high susceptibility were mainly distributed in rivers,both sides of the roads and areas where faults were developed;and the landslide susceptibility of the whole line was distributed in segments,with the most serious in the Shannan-Langxian and Bomi-Litang sections.The area under ROC curve of 1D CNN was 0.958,which could effectively evaluate the landslide susceptibility in this area.

关 键 词：一维卷积神经网络 滑坡敏感性评价 深度学习 川藏铁路 

分 类 号：P642.22[天文地球—工程地质学]