UM  > 科技學院  > 電腦及資訊科學系
Hyperspectral Image Classification Based on Spectral-Spatial One-Dimensional Manifold Embedding
Huiwu Luo1; Yuan Yan Tang1; Yulong Wang1; Jianzhong Wang2; Lina Yang3; Chunli Li1; Tingbo Hu1
2016-09-01
Source PublicationIEEE Transactions on Geoscience and Remote Sensing
ISSN01962892
Volume54Issue:9Pages:5319-5340
Abstract

A novel approach called Spectral-Spatial 1-D Manifold Embedding (SS1DME) is proposed in this paper for remotely sensed hyperspectral image (HSI) classification. This novel approach is based on a generalization of the recently developed smooth ordering model, which has gathered a great interest in the image processing area. In the proposed approach, first, we employ the spectral-spatial information-based affinity metric to learn the similarity of HSI pixels, where the contextual information is encoded into the affinity metric using spatial information. In our derived model, based on the obtained affinity metric, the created multiple 1-D manifold embeddings (1DMEs) consist of several different versions of 1DME of the same set of all HSI points. Since each 1DME of the data is a 1-D sequence, a label function on the data can be obtained by applying the simple 1-D signal processing tools (such as interpolation/regression). By collecting the predicted labels from these label functions, we build a subset of the current unlabeled points, on which the labels are correctly labeled with high confidence. Next, we add a proportion of the elements from this subset to the original labeled set to get the updated labeled set, which is used for the next running instance. Repeating this process for several loops, we get an extended labeled set, where the new members are correctly labeled by the label functions with much high confidence. Finally, we utilize the extended labeled set to build the target classifier for the whole HSI pixels. In the whole process, 1DME plays the role of learning data features from the given affinity metric. With the incrementation of learning features during iteration, the proposed scheme will gradually approximate the exact labels of all sample points. The proposed scheme is experimentally demonstrated using four real HSI data sets, exhibiting promising classification performance when compared with other recently introduced spatial analysis alternatives.

Keyword1-d Embedding High-dimensional Data Processing Hyperspectral Image (Hsi) Classification Manifold Learning (Ml) Smooth Interpolation Smooth Ordering Spatial Context Spectral-spatial Metric
DOIhttps://doi.org/10.1109/TGRS.2016.2560529
URLView the original
Language英语
WOS Research AreaGeochemistry & Geophysics ; Engineering ; Remote Sensing ; Imaging Science & Photographic Technology
WOS SubjectGeochemistry & Geophysics ; Engineering, Electrical & Electronic ; Remote Sensing ; Imaging Science & Photographic Technology
WOS IDWOS:000382689300024
PublisherIEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 445 HOES LANE, PISCATAWAY, NJ 08855-4141 USA
Fulltext Access
Citation statistics
Cited Times [WOS]:11   [WOS Record]     [Related Records in WOS]
Document TypeJournal article
CollectionDEPARTMENT OF COMPUTER AND INFORMATION SCIENCE
Corresponding AuthorHuiwu Luo; Yuan Yan Tang; Yulong Wang; Jianzhong Wang; Lina Yang; Chunli Li; Tingbo Hu
Affiliation1.Faculty of Science and Technology, University of Macau, Macau, China
2.Department of Mathematics and Statistics, Sam Houston State University, Huntsville, TX 77341 USA
3.School of Computer, Electronics and Information, Guangxi University, Nanning 530004, China
4.Sam Houston State University
5.Institute of Automation, College of Mechatronics and Automation, National University of Defense Technology, Changsha 410073, China
First Author AffilicationFaculty of Science and Technology
Corresponding Author AffilicationFaculty of Science and Technology
Recommended Citation
GB/T 7714
Huiwu Luo,Yuan Yan Tang,Yulong Wang,et al. Hyperspectral Image Classification Based on Spectral-Spatial One-Dimensional Manifold Embedding[J]. IEEE Transactions on Geoscience and Remote Sensing,2016,54(9):5319-5340.
APA Huiwu Luo.,Yuan Yan Tang.,Yulong Wang.,Jianzhong Wang.,Lina Yang.,...&Tingbo Hu.(2016).Hyperspectral Image Classification Based on Spectral-Spatial One-Dimensional Manifold Embedding.IEEE Transactions on Geoscience and Remote Sensing,54(9),5319-5340.
MLA Huiwu Luo,et al."Hyperspectral Image Classification Based on Spectral-Spatial One-Dimensional Manifold Embedding".IEEE Transactions on Geoscience and Remote Sensing 54.9(2016):5319-5340.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Huiwu Luo]'s Articles
[Yuan Yan Tang]'s Articles
[Yulong Wang]'s Articles
Baidu academic
Similar articles in Baidu academic
[Huiwu Luo]'s Articles
[Yuan Yan Tang]'s Articles
[Yulong Wang]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Huiwu Luo]'s Articles
[Yuan Yan Tang]'s Articles
[Yulong Wang]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.