文献类型：期刊文献

中文题名：GIS辅助下的基于数据挖掘的林型遥感分类方法研究

英文题名：Study on the GIS-aided remote sensing classification methods of forest land cover based upon date mining

作者：邓书斌[1] 武红敢[2] 江涛[1]

第一作者：邓书斌

机构：[1]山东科技大学,山东青岛266510;[2]中国林科院资源信息研究所,北京100091

年份：2008

卷号：33

期号：3

起止页码：121-123

中文期刊名：测绘科学

外文期刊名：Science of Surveying and Mapping

收录：CSTPCD;;北大核心:【北大核心2004】；CSCD:【CSCD2011_2012】；

基金：国家科技基础条件平台工作重点项目“森林灾害的监测、预警与管理系统平台(2003DIA6N007)”

语种：中文

中文关键词：数据挖掘;知识发现;分类规则;决策树;专家分类器

外文关键词：data mining; knowledge discovery classification rules decision treeexpert classifier

分类号：TP75

摘要：在植被分类领域,各类植被光谱特征非常相近,纹理特征比较单一,并且受物候影响较大,使得影像上林地信息的分类识别的成为一个较复杂的过程。本文利用GIS平台构建由影像、DEM等组成的多元空间数据库,将数据挖掘的方法引入林型遥感分类中,充分利用了现有的空间数据资源,减少了同物异谱、异物同谱对林型遥感分类的影响,增强了分类知识获取的自动化和智能化程度。实验结果表明,该方法具有简化知识获取过程和提高分类精度等优点。
In the field of vegetation classification, varieties of vegetation are similar in spectrum and texture characters and apt to be affected by the cli＇mate, which make the classification identification a complicated process. In this text we use GIS platform to construct a multi-dimension database made up of images, DEM, etc, introduce the methods of data mining into forest type RS classification, make full use of the spatial data resource available, to reduce the influence to the forest classification of the special situations which is that the same matter presents different spectrum and the same spectrum shows different matter, and improve the automatization and intelligentization of classification of knowledge acquisition. This method is proved by experiments to have got the excellency of predigesting the Classification of knowledge acquisition process and improving classification precise.