基于二元逻辑回归模型的新疆果子沟滑坡风险区划

发布时间：2018-02-20 01:19

本文关键词： RS 滑坡灾害频率比 Logistic回归模型易发性　出处：《山地学报》2017年02期 　论文类型：期刊论文

【摘要】：新疆果子沟区域是我国与中亚地区社会、经济、文化交流的重要通道以及我国石油、天然气等能源资源安全大通道,沿途滑坡灾害频发,威胁人类安全、影响社会、经济发展,但该区域针对滑坡灾害的研究程度较低,需借助滑坡灾害易发性分析与风险等级评估结果指导防灾减灾。本文基于GF-1号卫星影像数据进行滑坡解译,选取地层岩性、断层密度、坡度、坡向、地表高程、植被指数等6个评估因子,探讨运用GIS、RS技术及统计分析模型进行滑坡灾害易发性分析与建模。基于频率比法分析各因子敏感性,利用二元Logistic回归模型进行灾害易发性分析,将研究区滑坡灾害风险分为极低、低、中、高、极高五个等级。将模型计算结果与历史滑坡信息进行比较,并借助ROC Curve检验模型准确性,AUC为0.844,表明模型预测结果具有较高准确性,因此建立的分析模型可以满足新疆果子沟区域滑坡灾害分析与评估应用,研究成果可为研究区重大线型工程保护、边坡加固提供辅助决策支持。
[Abstract]:The Shougou region of Xinjiang is an important channel for social, economic and cultural exchanges between China and Central Asia, as well as a large channel for the safety of energy resources such as oil and natural gas in China. The frequent landslides along the way threaten human security and affect social and economic development. However, the research level of landslide disaster in this area is relatively low, so it is necessary to use the results of vulnerability analysis and risk grade evaluation of landslide disaster to guide disaster prevention and mitigation. Based on the satellite image data of GF-1, this paper interprets landslide, selects stratigraphic lithology, fault density, etc. In this paper, GISRS technique and statistical analysis model are used to analyze and model the vulnerability of landslide hazard. The sensitivity of each factor is analyzed based on the frequency ratio method, which includes six evaluation factors, such as slope, slope direction, surface elevation, vegetation index and so on. The risk of landslide hazard in the studied area is divided into five grades: extremely low, low, medium, high and very high by using the binary Logistic regression model. The results of the model are compared with the historical landslide information. With the help of ROC Curve, the accuracy of the model is 0.844, which indicates that the prediction result of the model has a high accuracy. Therefore, the established model can meet the needs of the analysis and evaluation of landslide disaster in Shegou region, Xinjiang. The research results can provide auxiliary decision support for important linear engineering protection and slope reinforcement in the study area.
【作者单位】：新疆大学资源与环境科学学院;新疆卫星应用中心;滁州学院;
【基金】：国家高分辨率对地观测系统重大专项项目(95-Y40B02-9001-13/15)~~
【分类号】：P642.22

【相似文献】