Class Separability Analysis and Classifier Comparison using Quad-polarization Radar Imagery

The potential of quad polarization radar data for the target discrimination has been analyzed. Quad polarization data of the RADARSAT-2 fine resolution mode has been utilized. Class separability analysis has been carried out on different polarization combinations using Transformed Divergence (TD) me...

Full description

Saved in:
Bibliographic Details
Published in:Journal of the Indian Society of Remote Sensing 2013-03, Vol.41 (1), p.177-182
Main Authors: Joseph, Manoj, Subramoniam, S. Rama, Srinivasan, K. S, Pathak, Suparn, Sharma, J. R
Format: Article
Language:English
Subjects:
Earth and Environmental Science
Earth Sciences
radar
remote sensing
Remote Sensing/Photogrammetry
Short Note
support vector machines
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The potential of quad polarization radar data for the target discrimination has been analyzed. Quad polarization data of the RADARSAT-2 fine resolution mode has been utilized. Class separability analysis has been carried out on different polarization combinations using Transformed Divergence (TD) method and it is observed that HH-HV/VH-VV polarization combination gives better class separability when compared to other polarization combinations. Classification has been carried out on the optimized polarization combination using Maximum likelihood (MLC) and Support Vector Machine (SVM) classifiers. It is observed that SVM classification gives better classification accuracy compared to MLC. Overall classification accuracy is 93.03% for SVM and 88.78% for MLC. Class separability and classification accuracy comparison results are presented.
ISSN:0255-660X
0974-3006
DOI:10.1007/s12524-011-0177-0