基于样本优化策略的滑坡易发性评价  

Evaluation of Landslide Susceptibility Based on Sample Optimization Strategy

作　　者：吴宏阳 周超[1,2] 梁鑫 王悦 袁鹏程 吴立星 WU Hongyang;ZHOU Chao;LIANG Xin;WANG Yue;YUAN Pengcheng;WU Lixing(School of Geography and Information Engineering,China University of Geosciences(Wuhan),Wuhan 430074,China;Research Center of Geohazard Monitoring and Warning in the Three Gorges Reservoir,Chongqing 404199,China;Faculty of Engineering,China University of Geosciences(Wuhan),Wuhan 430074,China)

机构地区：[1]中国地质大学(武汉)地理与信息工程学院,湖北武汉430074 [2]三峡库区地质灾害野外监测与预警示范中心,重庆404199 [3]中国地质大学(武汉)工程学院,湖北武汉430074

出　　处：《武汉大学学报（信息科学版）》2024年第8期1492-1502,共11页Geomatics and Information Science of Wuhan University

基　　金：国家自然科学基金(42371094,41907253)。

摘　　要：准确的易发性评价结果能够对滑坡带来的危险进行精准防控。样本优化是滑坡易发性评价的重要方法,可有效解决不平衡样本产生的决策边界偏移问题,提升滑坡易发性评价精度。以中国重庆市万州区东南区域为例,选取地层、土地利用、高程等10个影响因子构建滑坡易发性评价指标体系,应用频率比方法定量分析滑坡与指标之间的关系,在此基础上分别利用深度神经网络模型(deep neural networks, DNN)、过采样-深度神经网络模型(synthetic minority oversampling technique-DNN, SMOTE-DNN)、混合采样-深度神经网络耦合模型(one-class support vector machine-SMOTE-DNN, OS-DNN)、混合采样-深度神经网络-K均值聚类耦合模型(OS-DNN-K-means)进行滑坡易发性评价。结果表明,距道路距离、土地利用、地层是研究区滑坡发育的主要控制因子。精度评价结果发现OS-DNN-K-means(95.61%)和OS-DNN(91.16%)相较于模型SMOTE-DNN(87.97%)和DNN(81.40%)更能有效提高滑坡预测精度。通过混合采样和半监督分类进行样本优化能够有效解决研究区样本不平衡问题,为滑坡灾害空间预测提供新技术支撑。Objectives:Accurate susceptibility evaluation results can accurately prevent and control the dangers caused by landslides.Sample optimization is an important method for landslide susceptibility evaluation,which can effectively solve the problem of decision boundary offset generated by unbalanced samples and improve the accuracy of landslide susceptibility evaluation.Methods:Taking the southeast area of Wanzhou District of Chongqing,China as an example,ten influencing factors such as strata,land use and elevation were selected to construct a landslide susceptibility evaluation index system,and the relationship between landslide and the indices was quantitatively analyzed by frequency ratio method,and on this basis,deep neural network model(DNN),synthetic minority oversampling technique-DNN model(SMOTEDNN),one-class support vector machine-DNN coupling model(OS-DNN),and OS-DNN-K-means clustering coupling model(OS-DNN-K-means)were used to evaluate landslide susceptibility.Results:The results show that the distance from the road,land use and strata are the main control factors for landslide development in the study area.The accuracy evaluation results show that OS-DNN-K-means (95.61%) and OS-DNN (91.16%) could improve the landslide prediction accuracy more effectively com ‑pared with SMOTE-DNN (87.97%) and DNN (81.40%). Conclusions: Sample optimization throughmixed sampling and semi-supervised classification can effectively solve the problem of sample imbalance inthe study area, and provide new technical support for spatial prediction of landslide disasters.

关 键 词：滑坡 易发性建模 深度神经网络 混合采样 K均值聚类 样本优化策略 

分 类 号：P642.4[天文地球—工程地质学] P237[天文地球—地质矿产勘探]