Hyperspectral Image Classification Using Discrete Space Model and Support Vector Machines
Authors:
- 24 January 2017
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 14 (3) , 374-378
Abstract
In this letter, a novel method based on discrete space model (DSM) and support vector machines (SVMs) is proposed for hyperspectral image (HSI) classification. The DSM approach transforms continuous spectral signatures into discrete features and constructs a space model with the discrete features. Therefore, the classification capability of SVMs can be improved on account of the discrete feature space. Moreover, a composite kernel model is employed to take advantage of the spectral and spatial features among neighboring pixels. The proposed method is applied to real HSIs for classification. The experimental results confirm that the classification accuracy for the SVMs could be improved using the DSM method prior to classification.Keywords
Funding Information
- National Nature Science Foundation of China (60771065, 51378365)
This publication has 13 references indexed in Scilit:
- Novel classification method for remote sensing images based on information entropy discretization algorithm and vector space modelComputers & Geosciences, 2016
- Remotely Sensed Image Classification Using Sparse Representations of Morphological Attribute ProfilesIEEE Transactions on Geoscience and Remote Sensing, 2013
- Hyperspectral Image Classification Using Band Selection and Morphological ProfilesIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013
- Spectral–Spatial Classification of Hyperspectral Images Based on Hidden Markov Random FieldsIEEE Transactions on Geoscience and Remote Sensing, 2013
- Hyperspectral Image Classification via Kernel Sparse RepresentationIEEE Transactions on Geoscience and Remote Sensing, 2012
- A Spatial–Contextual Support Vector Machine for Remotely Sensed Image ClassificationIEEE Transactions on Geoscience and Remote Sensing, 2011
- A Composite Semisupervised SVM for Classification of Hyperspectral ImagesIEEE Geoscience and Remote Sensing Letters, 2009
- An Innovative Method to Classify Remote-Sensing Images Using Ant Colony OptimizationIEEE Transactions on Geoscience and Remote Sensing, 2008
- Investigation of the random forest framework for classification of hyperspectral dataIEEE Transactions on Geoscience and Remote Sensing, 2005
- Adaptive Bayesian contextual classification based on Markov random fieldsIEEE Transactions on Geoscience and Remote Sensing, 2002