Abstract

Classification of Hyperspectral Remote Sensing Images whit Support Vector Machines

Majid Khalifeh Gholi
Msc. Student of Remote Sensing
K.N. Toosi University of Technology, Iran
majid_khalifehgholi@yahoo.com

Abbas Alimohammadi
Assistant Professor
K.N. Toosi University of Technology, Iran
alimoh_abb@yahoo.com

M. Javad Valadan Zoej
Assistant Professor
K.N. Toosi University of Technology, Iran
valadanzouj@kntu.ac.ir

Barat Mojaradi
Ph.D Student
K.N. Toosi University of Technology, Iran
mojaradi@albors.kntu.ac.ir


Abstract :
Considering the Hughes phenomena in Hyperspectral data caused by limited trainings and high dimensionality of data, conventional classifiers does not work efficiently. Assuming the statistical properties of such data, the statistical classifiers which use the estimating probability functions does not work properly. To improve the efficiency of classification of hyperspectral data, avoiding the statistical properties, employing geometric classifiers is propounded. In this research, Support Vector Machines (SVMs) as method with geometrical concepts is studied. In order to evaluate this algorithm, the classification accuracy and the computational time is assessed. Experimental studies were carried out on the basis of hyperspectral images acquired by the Airborne Visible/Infrared Imaging Spectroradiometer (AVIRIS) sensor. Final results show how SVMs can improve the classification accuracy.