摘要
利用多时相的NOAA-AVHRR8km分辨率的遥感影像,以决策树分类器为基础,辅以数字化地形数据(DTM)、历史资料和野外实地调查资料等辅助分类数据,综合运用非监督分类和基于知识挖掘的信息提取技术,对中国东北地区20世纪80年代的地表覆被类型进行了分类,将研究区域最终划定为11种土地覆被类型,揭示了当时研究区域的土地覆被空间分异特征。
With the method based on decision tree classification, the time series Normalized Difference Index （ NDVI, from 1982 to 1986 with 8 km spatial resolution） data derived from NOAA - AVHRR were used to explore the surface land cover condition in the early 1980s. In order to increase the accuracy of classification, some ancillary data, such as DEM, the map of vegetation distribution of vegetation in China, historical data and field investigation data were used to provide more information. The adopted approach involved variable hierarchy tree according to spectrum reflect characteristics, land surface characteristics, phonological differences among vegetation types, leaf shape, etc. In the end, the surface land types were classified into 11 classes, which revealed the differentiation character of land cover space of this research area at that time.
出处
《东北林业大学学报》
CAS
CSCD
北大核心
2005年第5期38-40,共3页
Journal of Northeast Forestry University
基金
国家自然科学基金资助(30472714
30270254)。
