摘要
地处西南的渝北地区地表覆盖类型复杂、土地利用多元化,仅依赖于光谱特征的传统遥感信息提取方法难以获得较高的分类精度。利用决策树分类技术对渝北地区的TM遥感影像进行分类,除光谱信息外还结合地质、NDVI、PCI等多源数据进行实验。结果表明,总精度和Kappa系数分别为88.42%和0.854 7,较传统的监督分类和仅依赖于光谱特征的决策树分类方法有较大提高,这也表明基于多源数据的决策树分类技术对地表覆盖复杂地区的遥感影像分类比较适用,是遥感信息提取的一种有效手段。
Yubei located in the southwest of China where covered varied landforms and diversified land use.Traditional remote sensing information extraction methods cannot get high accuracy of classification.This paper classified the TM images of Yubei by using classification technology of decision tree.It considered geology,NDVI and PCI besides spectral data in the procession of classification.The classification results showed that the total classification accuracy was 88.42%,and the Kappa coefficient was 0.854 7.The accuracy of classification based on multivariate data was higher than the method only based on spectral data and traditional supervised classification.It approved that the classification technology of decision tree based on multivariate data was applicable to the classification of remote sensing image of rolling country.
出处
《地理空间信息》
2012年第2期110-112,182,共3页
Geospatial Information
基金
国家自然科学基金资助项目(40972225)
成都理工大学青年科学基金资助项目(2010QJ08)
成都理工大学高层次人才科研启动基金资助项目(HJ0070)
关键词
决策树
遥感
渝北
地表覆盖
decision tree,remote sensing,Yubei district,land cover
