Share Email Print

Proceedings Paper

Unsupervised optimal fuzzy clustering and Markov segmentation of polarimetric imaging
Author(s): Safwan El Assad; Ali Saad; Dominique Barba
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

This paper presents a method for unsupervised segmentation of polarimetric SAR data into classes of homogeneous microwave backscatter characteristics. Clustering of polarimetric backscatter are obtained either by the CMF-NSO or be SEM algorithm. These algorithms carry out the classification without a priori assumptions on the number of classes in the data set. Assessment of cluster validity is based on performance measures using hypervolume V or CS function criteria. The later measures the overall average compactness and separation of a fuzzy-partition. The CMF-NSO algorithm performs well in situations of large variability of cluster shapes and densities. Given the clusters of polarimetric backscatter, the entire image is segmented using a MAP estimation. Implementation of the MAP technique is accomplished by an ICM algorithm. Results, using fully polarimetric SAR forest data, obtained by the CMF-NSO following by the ICM algorithm with a K-distribution model are quite satisfactory.

Paper Details

Date Published: 17 December 1996
PDF: 12 pages
Proc. SPIE 2958, Microwave Sensing and Synthetic Aperture Radar, (17 December 1996); doi: 10.1117/12.262691
Show Author Affiliations
Safwan El Assad, IRESTE (France)
Ali Saad, IRESTE (France)
Dominique Barba, IRESTE (France)

Published in SPIE Proceedings Vol. 2958:
Microwave Sensing and Synthetic Aperture Radar
Giorgio Franceschetti; Christopher John Oliver; Franco S. Rubertone; Shahram Tajbakhsh, Editor(s)

© SPIE. Terms of Use
Back to Top