陕北黄土高原地区土地利用/覆被分类及验证  

Land use and land cover classification and verification in Northern Shaanxi Loess Plateau

作　　者：宋富强[1,2] 康慕谊[2] 郑壮丽[3,4] 王令超[1] 王国强[1] 冯德显[1] 

机构地区：[1]河南省科学院地理研究所,郑州450052 [2]地表过程与资源生态国家重点实验室,北京师范大学资源学院,北京100875 [3]中国科学院华南植物园,广州510650 [4]中国科学院研究生院,北京100049

出　　处：《农业工程学报》2011年第3期316-324,I0009,I0010,共11页Transactions of the Chinese Society of Agricultural Engineering

基　　金：国家重点基础研究发展计划项目(2007CB106807);国家自然科学基金项目(40671065)

摘　　要：陕北地区自退耕还林工程实施以来,区域土地覆被发生巨大变化,退耕后土地利用/覆被分类有利于实现退耕前后土地利用/覆被动态监测,评价退耕还林工程进展程度。该文在前人分类系统的基础上,通过对研究区不同土地覆被类型分布特征的实地调查,首先构建了适合陕北地区的土地覆被分类系统;其次依照上述分类系统,以MODIS/NDVI数据、DEM数据和陆面温度数据为依据,采用决策树分类方法对2008年陕北地区土地覆被进行分类;最后采用野外调查数据和Google Earth高分辨率遥感数据,利用误差矩阵的方法对分类结果的整体精度和典型区域不同土地覆被类型的区分精度进行验证。结果表明,基于上述方法的土地覆被分类不仅反映了土地覆被分布的整体情况,而且不同的土地覆被类型之间也得到较好的区分。Since the implementation of the Grain for Green project in 1999,land use and land cover had changed apparently in Northern Shaanxi Loess Plateau.Land cover classification after the implementation of Grain for Green project,which was important in monitoring the land use and land cover change（LUCC） and assessing the project.Based on the previous classification system and the field survey of the different land cover distribution,a suitable land cover classification system for Northern Shaanxi was constructed firstly,and then according to MODIS/NDVI,DEM and land surface temperature data,the land cover classification in the Northern Shaanxi Loess Plateau was finished by decision tree method.Finally,using the field survey data and Google Earth high-resolution remote sensing data,the classification accuracy of the whole area and different land cover types was verified by error matrix method.The results showed that land cover classification of this study not only reflect the whole distribution of land use and land cover,but also the distinction between different land cover types.

关 键 词：遥感 分类 土地利用/覆被变化 陕北黄土高原 退耕还林 MODIS/NDVI 

分 类 号：F301.24[经济管理—产业经济]