
第21卷第11期 2013年11月
1004-924X(2013)11-2922-09
文章编号
光学精密工程
Optics and Precision Engineering
Vol. 21No. 11
Nov,2013
基于稀疏鉴别嵌入的高光谱遥感影像分类
黄鸿1,2*,杨媚,张满菊3
(1.重庆大学光电技术及系统教育部重点实验室,重庆400030; 2.重庆川仪自动化股份有限公司技术中心,重庆401121;
3.酒泉卫星发射中心,甘肃酒泉735300)
摘要:稀疏保持投影(SPP)是一种基于1,图的新型降维算法,它利用样本间的稀疏重构关系建图,但是SPP为非监督算法,分类效果受到限制。针对此间题,本文提出了一种新的稀疏流形学习算法一稀疏鉴别嵌人(SDE)。该算法在利用样
人空间中尽可能地分散开,SDE通过结合数据稀疏性及类间流形结构的优点,不仅保留样本间的稀疏重构关系,面且通过引人训练样本的类别信息实现稀疏鉴别特征提取,更有利于分类。在Urban和WashingtonDCMall数据集上的实验结果表明;SDE算法比其他算法的分类性能有明显的提升,在每类随机选取16个训练样本的情况下,SDE算法的分类精度分别达到了73.47%和98.35%。
关键词:高光谱遮感影像;维数约简;静疏表示;流形学习:薪疏甚别嵌入
中图分类号:TP751.1
文献标识码:A
doi: 10, 3788/OPE, 20132111. 2922
HyperspectralremotesensingimageclassificationbasedonSDE
HUANG Hong-2-, YANG Mei', ZHANG Man-ju
(l.Key Laboratory of Opto-electronic Technology &. Systems of the Ministry of Education,
Chongqing Uniersity,Chongqing 400030,Chind;
2.The Technical Center of Chongqing Chuanyi Automation Co.,Ltd.,Chongqing 401121,China;
3.JiuquanSatelliteLaunchCenter,Jiuquan735300,China) +Correspondingauthor,E-mail:hhuang.cqu@gmail.com
Abstract: Sparsity Preserving Projection(SPP) is a new algorithm for reducing dimensions of dataset based on a weighted graph( L,-Graph), which reconstructs the weighted graph by the sparse relation ship of train samples. However, SPP is an unsupervised learning method essentially, and it doesn't employ any prior knowledge of class to extract identification features. For this issue, a novel algo-rithm, Sparsity Discriminant Embedding (SDE) is proposed. Unlike SPP, the SDE adopts the class information of train samples when it constructs weighted graph of sparse reconstruction relationship. The projection matrix of the SDE is obtained via optimizing objective function and making different kinds of data points separate in the low-dimensional embedding space via a projection. By combining
收稿日期:2013-04-01;修订日期:2013-04-24.
基金项目:国家自然科学基金资助项日(No.61101168);中国博士后科学基金资助项目(No.2012M511906);重庆市
博士后科研基金特别资助项目(No.XM2012001)