Extended random walkers for hyperspectral image classification

Xudong Kang; Shutao Li; Meixiu Li; Jon Atli Benediktsson

doi:10.1109/IGARSS.2014.6946727

Extended random walkers for hyperspectral image classification

Xudong Kang
, Shutao Li
, Meixiu Li
, Jon Atli Benediktsson

University of Iceland

University of Iceland

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

A novel spectral-spatial hyperspectral image classification is proposed based on extended random walkers. First, a widely used pixel-wise classifier, i.e., the support vector machine (SVM), is adopted to obtain probability maps for a hyper-psectral image, which measure the probabilities that a pixel belongs to different classes. Then, the initial probabilities are optimized with the extended random walkers. Finally, by assigning each pixel with the label for which the greatest probability is obtained, the classification result is obtained. Experiments show the outstanding performance of the proposed method in terms of classification accuracy especially when the number of training samples is relatively small.

Original language	English
Title of host publication	International Geoscience and Remote Sensing Symposium (IGARSS)
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	1520-1523
Number of pages	4
ISBN (Electronic)	9781479957750
DOIs	https://doi.org/10.1109/IGARSS.2014.6946727
Publication status	Published - 4 Nov 2014
Event	Joint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014 - Quebec City, Canada Duration: 13 Jul 2014 → 18 Jul 2014

Publication series

Name	International Geoscience and Remote Sensing Symposium (IGARSS)

Conference

Conference	Joint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014
Country/Territory	Canada
City	Quebec City
Period	13/07/14 → 18/07/14

Bibliographical note

Other keywords

Extended random walkers
hyperspectral image
optimization
spectral-spatial classification

Access to Document

10.1109/IGARSS.2014.6946727

Cite this

Kang, X., Li, S., Li, M., & Benediktsson, J. A. (2014). Extended random walkers for hyperspectral image classification. In International Geoscience and Remote Sensing Symposium (IGARSS) (pp. 1520-1523). Article 6946727 (International Geoscience and Remote Sensing Symposium (IGARSS)). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IGARSS.2014.6946727

@inproceedings{8974a92686724c62a3b0689f2289c487,

title = "Extended random walkers for hyperspectral image classification",

abstract = "A novel spectral-spatial hyperspectral image classification is proposed based on extended random walkers. First, a widely used pixel-wise classifier, i.e., the support vector machine (SVM), is adopted to obtain probability maps for a hyper-psectral image, which measure the probabilities that a pixel belongs to different classes. Then, the initial probabilities are optimized with the extended random walkers. Finally, by assigning each pixel with the label for which the greatest probability is obtained, the classification result is obtained. Experiments show the outstanding performance of the proposed method in terms of classification accuracy especially when the number of training samples is relatively small.",

keywords = "Extended random walkers, hyperspectral image, optimization, spectral-spatial classification",

author = "Xudong Kang and Shutao Li and Meixiu Li and Benediktsson, \{Jon Atli\}",

note = "Publisher Copyright: {\textcopyright} 2014 IEEE.; Joint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014 ; Conference date: 13-07-2014 Through 18-07-2014",

year = "2014",

month = nov,

day = "4",

doi = "10.1109/IGARSS.2014.6946727",

language = "English",

series = "International Geoscience and Remote Sensing Symposium (IGARSS)",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "1520--1523",

booktitle = "International Geoscience and Remote Sensing Symposium (IGARSS)",

address = "United States",

}

Kang, X, Li, S, Li, M & Benediktsson, JA 2014, Extended random walkers for hyperspectral image classification. in International Geoscience and Remote Sensing Symposium (IGARSS)., 6946727, International Geoscience and Remote Sensing Symposium (IGARSS), Institute of Electrical and Electronics Engineers Inc., pp. 1520-1523, Joint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014, Quebec City, Canada, 13/07/14. https://doi.org/10.1109/IGARSS.2014.6946727

Extended random walkers for hyperspectral image classification. / Kang, Xudong; Li, Shutao; Li, Meixiu et al.
International Geoscience and Remote Sensing Symposium (IGARSS). Institute of Electrical and Electronics Engineers Inc., 2014. p. 1520-1523 6946727 (International Geoscience and Remote Sensing Symposium (IGARSS)).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Extended random walkers for hyperspectral image classification

AU - Kang, Xudong

AU - Li, Shutao

AU - Li, Meixiu

AU - Benediktsson, Jon Atli

PY - 2014/11/4

Y1 - 2014/11/4

N2 - A novel spectral-spatial hyperspectral image classification is proposed based on extended random walkers. First, a widely used pixel-wise classifier, i.e., the support vector machine (SVM), is adopted to obtain probability maps for a hyper-psectral image, which measure the probabilities that a pixel belongs to different classes. Then, the initial probabilities are optimized with the extended random walkers. Finally, by assigning each pixel with the label for which the greatest probability is obtained, the classification result is obtained. Experiments show the outstanding performance of the proposed method in terms of classification accuracy especially when the number of training samples is relatively small.

AB - A novel spectral-spatial hyperspectral image classification is proposed based on extended random walkers. First, a widely used pixel-wise classifier, i.e., the support vector machine (SVM), is adopted to obtain probability maps for a hyper-psectral image, which measure the probabilities that a pixel belongs to different classes. Then, the initial probabilities are optimized with the extended random walkers. Finally, by assigning each pixel with the label for which the greatest probability is obtained, the classification result is obtained. Experiments show the outstanding performance of the proposed method in terms of classification accuracy especially when the number of training samples is relatively small.

KW - Extended random walkers

KW - hyperspectral image

KW - optimization

KW - spectral-spatial classification

UR - https://www.scopus.com/pages/publications/84911396548

U2 - 10.1109/IGARSS.2014.6946727

DO - 10.1109/IGARSS.2014.6946727

M3 - Conference contribution

T3 - International Geoscience and Remote Sensing Symposium (IGARSS)

SP - 1520

EP - 1523

BT - International Geoscience and Remote Sensing Symposium (IGARSS)

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - Joint 2014 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2014 and the 35th Canadian Symposium on Remote Sensing, CSRS 2014

Y2 - 13 July 2014 through 18 July 2014

ER -

Extended random walkers for hyperspectral image classification

Abstract

Publication series

Conference

Bibliographical note

Other keywords

Access to Document

Fingerprint

Cite this