| 基于拉普拉斯特征映射高光谱遥感影像降维及其分类 |
| |
| 引用本文： | 钱进,邓喀中,范洪冬. 基于拉普拉斯特征映射高光谱遥感影像降维及其分类[J]. 遥感信息, 2012, 27(5): 3-7 |
| |
| 作者姓名： | 钱进 邓喀中 范洪冬 |
| |
| 作者单位： | 江苏省资源环境信息工程重点实验室,徐州221116;中国矿业大学国土环境与灾害监测国家测绘局重点实验室,徐州221116 |
| |
| 基金项目： | 高等学校博士学科点专项科研基金资助课题(20090095110002);国家自然科学基金(41071273) |
| |
| 摘 要： | 在进行高光谱遥感影像监督分类过程中,结合高光谱数据非线性的特点和流形学习强大的非线性处理能力,提出一种基于拉普拉斯特征映射(LE)降维和最佳指数法(OIF)波段组合选择训练样本进行SVM分类的策略,首先对高光谱遥感影像波段进行优化,利用拉普拉斯特征映射法(LE)对波段优选后的影像进行降维,利用OIF选择波段组合叠加进行训练样本选择。在此基础上采用支持向量机(SVM)进行分类处理,取得了优于PCA的效果。实验证明了流形学习是一种行之有效的高光谱遥感数据特征提取方法。
|
| 关 键 词： | 高光谱遥感 降维 流行学习 分类 拉普拉斯特征映射 |
Dimensionality Reduction and Classification for Hyperspectral Remote Sensing Imagery Based on Laplacian Eigenmap |
| |
| QIAN Jin , DENG Ka-zhong , FAN Hong-dong. Dimensionality Reduction and Classification for Hyperspectral Remote Sensing Imagery Based on Laplacian Eigenmap[J]. Remote Sensing Information, 2012, 27(5): 3-7 |
| |
| Authors: | QIAN Jin DENG Ka-zhong FAN Hong-dong |
| |
| Affiliation: | ①,②(① Jiangsu Key Laboratory of Resources and Environmental Information Engineering,XuZhou 221116; ② Key Laboratory for Land Resources and Disaster Monitoring of SBSM,China University of Mining and Technology,XuZhou 221116) |
| |
| Abstract: | In the supervised classification for hyperspectral imagery data,considering the non-linear characteristics of hyperspectral remote sensing data and the strong non-linear capacity of manifold learning,this paper proposed a new strategy based on Laplacian Eigenmap and Optimum Index Factor.First optimize the bands of hyperspectral data and reduce its dimensionality by LE,next choose the train samples based on the band selection by OIF.And then,get a better SVM result than PCA.Experiments show that manifold learning is effective for dimensionality reduction and feature extraction for hyperspectral remote sensing imagery. |
| |
| Keywords: | hyperspectral remote sensing dimensionality reduction manifold learning classification Laplacian Eigenmap |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
