摘要
通过对塔里木河下游MISR卫星多角度观测数据的不同组合构建多角度数据集,探索多角度观测与传统垂直观测对土地利用覆被遥感制图效果的影响,分别使用SVM(支持向量机)与传统的MLC(最大似然分类法)作为分类器,对分类后得到的混淆矩阵进行分析。结论证实:无论是使用传统的MLC还是SVM作为分类器,多角度观测都取得比垂直观测更高的总体分类精度;MISR近红外波段虽然分辨率较低,但依然含有丰富的信息,对地表覆被的分类有重要影响;无论使用哪一数据集,SVM法都能获得更高的分类精度;不同相机对分类结果的影响各不相同,其中C、D相机的作用更重要。
MISR multi-angular data have been built through 9 cameras combination,and the infulence on land use and land cover mapping by multi-angular observing and traditional nadir approach has been explored in the lower Tarim River.In addition,SVM（support vector machine）and conventional MLC（maximum likelihood classification）were respectively implemented to observe the differentiation of Confusion Matrix.The findings are presented as follows：Multi-angular oberservation achieved higher classification accuracy compared to nadir approach no matter MLC or SVM classifiers being used;Although the lower resolution,MISR obtains abundent information,and thus has a greatimpact on the classification of vegetation;the classification by SVM shows a higher accuracy than that by MLC no matter which data set is used;different cameras lead to different results,but camera C and D paly more important roles than the others.
出处
《吉林大学学报（地球科学版）》
EI
CAS
CSCD
北大核心
2016年第2期617-626,共10页
Journal of Jilin University：Earth Science Edition
基金
国家自然科学基金项目(41261051)
新疆维吾尔自治区重点实验室开放基金(XJDX0909-2010-08)~~
关键词
MISR
支持向量机
最大似然分类法
塔里木河下游
土地利用覆被
MISR
support vector machine
maximum lickelihood classification
lower Tarim River
land use and land cover
