A framework to estimate natural classes in remotely sensed data

Author

Prakash, H. N Srikanta ; Nagabhushan, P. ; Gowda, K. Chidananda

Author_Institution

Dept. of Comput. Sci. & Eng., S.J. Coll. of Eng., Mysore, India

fYear

1996

fDate

18-20 Nov 1996

Firstpage

264

Lastpage

266

Abstract

This paper presents a new approach to estimating the natural classes in remotely sensed data. The cluster index (CI) based on the mean difference index (MDI) concept for identifying the natural classes is proposed. This concept can be used in any classification procedures. In this paper, MDI uses the traditional symbolic ISODATA clustering algorithm to obtain the natural classes. The experimental results of multispectral IRS (Indian Remote Sensing) satellite data covering Mysore city, Karnataka State, India is encouraging. This endeavor is expected to open a new avenue in the area of CI for remotely sensed data

Keywords

data analysis; geophysical techniques; geophysics computing; pattern classification; remote sensing; symbol manipulation; Indian Remote Sensing; classification; cluster index; experimental results; mean difference index; multispectral IRS; natural class estimation; remotely sensed data; satellite data; symbolic ISODATA clustering algorithm; Australia; Cities and towns; Clustering algorithms; Clustering methods; Data analysis; Educational institutions; Intelligent systems; Merging; Remote sensing; Satellites;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Information Systems, 1996., Australian and New Zealand Conference on

Conference_Location

Adelaide, SA

Print_ISBN

0-7803-3667-4

Type

conf

DOI

10.1109/ANZIIS.1996.573953

Filename

573953