基于MODIS数据的松花江流域土地利用信息提取研究  

Land Use Information Extraction in the Basin of Songhua River by Using MODIS Data

作　　者：王亚楠[1] 陈圣波[1] 逄超[1] 陶玉龙[1] 

机构地区：[1]吉林大学地球探测科学与技术学院,吉林长春130026

出　　处：《安徽农业科学》2010年第6期3300-3302,3312,共4页Journal of Anhui Agricultural Sciences

摘　　要：以松花江流域为研究区域,利用单时相MODIS数据进行大范围区域土地利用信息提取;结合地表反射率、植被指数(NDVI)、土壤湿度指数(NDMI)、归一化水体指数(NDWI)等特征变量试验决策树分类法,与传统分类方法相比较,分析松花江流域土地利用变化情况。结果表明:决策树法优于传统的最大似然法,分类精度分别为83.48%和75.51%;2001～2006年,整体上松花江流域耕草地面积增加,林地面积减少;中国部分,耕草地面积增加,林地面积减少,而俄罗斯部分耕草地面积减少,林地面积增加。Taking Songhua River Basin as the research area,single-temporal MODIS data were used to extract land use information of large areas.Combined with the characteristic variables of surface albedo,vegetation index（NDVI）,soil moisture index（NDMI）,water index（NDWI） and so on,the test was made between maximum likelihood classifier and decision tree classifier for analyzing the land use change situations of Songhua River Basin.The results showed that the decision tree classifier was better than conventional maximum likelihood classifier and the classification accuracies were 83.48% and 75.51% respectively.The arable and grass land area of Songhua River Basin increased and woodland area decreased on the whole from 2001 to 2006.In China,arable and grass land area increased and woodland area decreased.In Russia,arable and grass land area decreased and woodland area increased.

关 键 词：MODIS 松花江 土地利用 特征变量 决策树 

分 类 号：F301.2[经济管理—产业经济]