Multiple Classifier Systems for Hyperspectral Remote Sensing Data Classification
- PDF / 257,149 Bytes
- 6 Pages / 595.276 x 790.866 pts Page_size
- 25 Downloads / 225 Views
SHORT NOTE
Multiple Classifier Systems for Hyperspectral Remote Sensing Data Classification Iman Khosravi & Majid Mohammad-Beigi
Received: 15 April 2013 / Accepted: 12 September 2013 # Indian Society of Remote Sensing 2013
Abstract One of the most widely used outputs of remote sensing technology is Hyperspectral image. This large amount of information can increase classification accuracy. But at the same time, conventional classification techniques are facing the problem of statistical estimation in high-dimensional space. Recently in remote sensing, support vector machines (SVMs) have shown very suitable performance in classifying high dimensionality problem. Another strategy that has recently been used in remote sensing is multiple classifier system (MCS). It can also improve classification accuracy by combining different classifier methods or by a diversity of the same classifier. This paper aims to classify a Hyperspectral data using the most common methods of multiple classifier systems i.e. adaboost and bagging and a MCS based on SVM. The data used in the paper is an AVIRIS data with 224 spectral bands. The final results show the high capability of SVMs and MCSs in classifying high dimensionality data.
Keywords Multiple classifier system . Support vector machine . Hyperspectral data classification . Correlation-based feature selection
I. Khosravi (*) Department of Surveying Engineering, Faculty of Engineering, University of Isfahan, Isfahan, I.R., Iran e-mail: [email protected] I. Khosravi e-mail: [email protected] M. Mohammad-Beigi Department of Biomedical Engineering, Faculty of Engineering, University of Isfahan, Isfahan, I.R., Iran e-mail: [email protected]
Introduction Overview One of the widely utilized outputs of remote sensing technology is the Hyperspectral image. This data covers a wide spectral range from visible to short-wavelength infrared at a large number of spectral channels. In a classification task, whatever the dimensionality of the data increases, the capability of detecting different classes is increased (Fauvel et al. 2008). However, the high number of features and inadequate number of training samples can decrease the accuracy of classification. This problem is known as Hughes phenomenon (Fauvel et al. 2006). In this case, conventional statistical algorithms such as maximum likelihood cannot produce appropriate results and it is necessary to consider other methods (Fauvel et al. 2006). Up to now, several different classification algorithms have been proposed for solving this problem. Recently in remote sensing, it has been shown that support vector machines (SVMs) has very suitable performance in classifying high dimensional data (e.g. (Camps-Valls and Bruzzone 2005; Ceamanos et al. 2010; Fauvel et al. 2006, 2008; Pal and Mather 2004). The first use of SVMs for classifying remotely sensed images had acceptable results (Gualtieri and Chettri 2000). Pal and Mather (2004) demonstrated that SVMs have high capability of generalization and they do not have the impact o
Data Loading...
