基于支撑向量机概率输出的高光谱影像混合像元分解

被引:14
作者
吴波
张良培
李平湘
机构
[1] 武汉大学测绘遥感信息工程国家重点实验室
关键词
支撑向量机; 多类; 后验概率; 像元分解; 高光谱;
D O I
10.13203/j.whugis2006.01.014
中图分类号
TP75 [遥感图像的解译、识别与处理];
学科分类号
081002 ;
摘要
提出利用支撑向量机(SVM)后验概率来分解高光谱影像的混合像元,通过支撑向量机的输出值转化为两两配对的后验概率,再由两两配对的概率值求得多类后验概率,并以像元所属类别的后验概率作为地物的组分信息。实验结果表明,该方法能较好地估计出混合像元的组分比。
引用
收藏
页码:51 / 54
页数:4
相关论文
共 7 条
[1]  
A Relative Evaluation ofMulticlass Image Classification by Support VectorMachines. Foody G M,Mathur A. IEEE Transactions on Geoscience andRemote Sensing . 2004
[2]  
Classification of Hyper-spectral Remote Sensing Images with Support Vec-tor Machines. Melgani F,Bruzzone L. IEEE Transactions on Geoscienceand Remote Sensing . 2004
[3]  
On the Use of Prior and Posterior Informa-tion in the Subpixel Proportion Problem. Eric D. IEEE Transactions on Geoscience and Remote Sensing . 2003
[4]  
An Overview of Statistical LearningTheory. Vapnik V. IEEE Transactions on Neural Net-works . 1999
[5]  
Classification by PairwiseCoupling. Hastie T,Tibshirani R. The Annals of Statistics . 1998
[6]  
Relating the Land Cover Compositionof Mixed Pixels of Artificial Neural Network Classi-fication Output. Foody G M. Photogrammtry Engineeringand Remote Sensing . 1996
[7]  
Relating the Land Cover Compositionof Mixed Pixels of Artificial Neural Network Classi-fication Output. Foody G M. Photogrammtry Engineeringand Remote Sensing . 1996