摘要
为实现简便、快速、准确地获取大尺度范围的土地覆盖信息,本文充分利用ETM+数据的多光谱特征、DEM数字高程信息和坡度、坡向等地学相关知识,结合NDVI、NDWI、SAVI、NDBI等各类指数,构建适用于研究区土地覆盖信息提取的决策树模型,并验证其精度。结果表明,该模型能够更好地适用于土地覆盖信息提取,总体分类精度达到86.49%,Kappa系数0.8367。
To get a lot of information from land cover extraction at a convenience, celerity and accuracy, this paper fully used the characteristics of ETM ＋ multispectral data, the DEM elevation information and the relevant knowledge about a slope,orientation etc. combining with each index of NDVI, NDWI, SAVI, NDBI to establish the decision tree model of land cover information extraction and verify its precision. The result showed that it was suitable enough for the extraction of land cover information and its precision wet up to 86.49%, the Kappa coefficient was 0.8367.
出处
《山东农业大学学报（自然科学版）》
CSCD
2016年第3期372-377,共6页
Journal of Shandong Agricultural University：Natural Science Edition
基金
山东省优秀中青年科学家奖励基金(2011BSB01500)
关键词
土地覆盖
ETM+
决策树分类
信息提取
Soil cover
ETM＋
decision tree classification
information extraction
