摘要
对澜沧江流域山区典型试验样区遥感数据运用AHP递阶层次结构 ,将土地覆盖类别分成若干层次。结合特征选取与采用多种分类算法组合 ,先进行类间易于区别的大类别的分类信息提取处理 ,得到一层次的分类结果 ,再基此对各分类结果探索进一步的分类处理 ,获得第二层次的分类结果。如此进行 ,直至分出所有确定类别。试验结果表明 ,该分类组织较之传统基于一次特征选取所进行的单级分类技术组织实施 ,具有构思科学合理 ,操作简单可行。监测结果评价精度可满足澜沧江流域综合开发 ,达到获取反映山地生态主要覆盖类别的遥感监测技术要求。
The effect of organization of classification process is one of the important contents in land cover remote monitoring in mountain area.The traditional classification approaches were performed in a single stage and have many problems.In this paper,the multi stage classification approach is performed.This is an iterative approach in which the types are classified one or two of a time.In the iterative classification process,the bands that can best determinte the cover types would be selected.Results on study of typical area indicate that the iterative multi stage classification approach was significantly better than the single stage classification approach.This organizational methodology for classification is feasible and reliable in Lancangjiang River basin by staellite imagery remote sensing monitoring.
出处
《资源科学》
CSSCI
CSCD
北大核心
2000年第5期78-82,共5页
Resources Science
基金
国家自然科学基金!"九五"重点项目 ( 4 963 10 2 0 )
云南省自然科学基金!"九五"重点项目( 96D0 12Z)资助
关键词
澜沧江
土地覆盖
遥感监测
分类组织
生态监测
Lancangjiang River
Land cover
Remote sensing monitoring
Classification organization
