基于粗糙集理论滑坡影响因子评价研究——以三峡库区青干河流域为例  

Assessment of impact factors for landslides based on rough sets theory:A case study on the Qingganhe River of the Three Gorges area

作　　者：刘吉平[1] 刘汉青[1] 曾忠平[2] 汪华斌[3] 

机构地区：[1]武汉大学资源环境学院,武汉430079 [2]华中科技大学公共管理学院,武汉430079 [3]华中科技大学土木工程与力学学院,武汉430079

出　　处：《水文地质工程地质》2010年第5期118-122,132,共6页Hydrogeology & Engineering Geology

基　　金：自然科学基金资助项目(40772170)

摘　　要：运用地理信息系统和遥感技术,从LandsatTM遥感图像获取了一系列滑坡及其影响因子数据。结合粗糙集理论,对三峡库区秭归县青干河流域滑坡发生的影响因子进行了分析,提取了一组基本的滑坡影响条件属性因子,并导出了基于该因子集合的判断滑坡与非滑坡的规则集。研究结果表明,所选择的坡度、高程、斜坡类型、植被指数和岩石地层单元等5个条件属性因子对滑坡是重要的影响因素(核);由粗糙集生成的对预测滑坡相对较有价值的11条决策规则中,3条主要决策规则可作为滑坡影响因子的评价规则。Landslides can be affected by a series of environmental factors.The present study aimed to analyze landslide susceptibility in the Qingganhe River of Three Gorges area in China.Spatial information techniques,including RS and GIS with an innovative data mining approach,namely,the rough sets theory,are incorporated to the study for landslide factors analysis.In a case study,the cores（i.e.,slope,elevation,slope type,vegetation index,and rock type） and reductions of the data attributes were obtained after data mining using the rough sets theory.Rules of all the impact factors,which can make contributions to landslide occurrence,were generated on the basis of landslide knowledge.All of the 11 rules were classified into straight rules and approximate rules,while three main rules were extracted to be the decision-making rules for landslide predictions in view of importance.

关 键 词：滑坡 数据挖掘 粗糙集理论 空间信息 

分 类 号：P642.22[天文地球—工程地质学]