FCM and HCA performance analysis for crop type classification of SAR imagery

Author

Martín, Maite Trujillo San ; Sadki, Mustapha

Author_Institution

Dept. of Electron. & Comput. Eng., Brunel Univ., Uxbridge, UK

Volume

4

fYear

2004

fDate

20-24 Sept. 2004

Firstpage

2692

Abstract

In this study, we investigate the classification performance of two clustering algorithms, the fuzzy C-means (FCM) and hierarchical clustering analysis (HCA) algorithms applied to crop type classification of high-resolution airborne synthetic aperture radar (SAR) imagery based on Haralick and autocorrelation textural features. The contribution of the different polarization channels toward the overall classification of different cluster regions are also analyzed as well as the influence in the election of the optimum parameters for wavelet image enhancement.

Keywords

crops; geophysical signal processing; image classification; image texture; pattern clustering; radar imaging; remote sensing by radar; synthetic aperture radar; vegetation mapping; wavelet transforms; Haralick; SAR imagery; airborne synthetic aperture radar; autocorrelation textural feature; clustering algorithm; crop type classification; fuzzy C-means; hierarchical clustering analysis; high-resolution imagery; image classification; polarization channel; wavelet image enhancement; Algorithm design and analysis; Autocorrelation; Clustering algorithms; Crops; Image analysis; Image texture analysis; Nominations and elections; Performance analysis; Polarization; Synthetic aperture radar;

fLanguage

English

Publisher

ieee

Conference_Titel

Geoscience and Remote Sensing Symposium, 2004. IGARSS '04. Proceedings. 2004 IEEE International

Print_ISBN

0-7803-8742-2

Type

conf

DOI

10.1109/IGARSS.2004.1369855

Filename

1369855