Share Email Print
cover

Proceedings Paper

Characterization of urban areas using principal component analysis from multitemporal ERS coherence imagery
Author(s): Attilio Fanelli; Maurizio Santoro
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

C-band SAR interferometry using ERS data has been shown to be potential for urban areas studies. This work illustrates the application of Principal Components Analysis (PCA) to a multi-temporal set of ERS coherence images to detect urban areas and their features. In particular Principal Component Transformation was applied on sets of one-day and long-term coherence images for urban mapping applications in the area of Naples, Italy. Two main classes, urban and non-urban, which then included two classes each, were considered in this study. Dense built-up areas and residential areas formed the urban class. Water bodies and vegetated areas (fields and woods) were grouped in the non-urban class. The first principal component was found to be more suitable than higher order components for detection of urban areas. Moreover, a simple algorithm based on distance between the first principal component of a pixel and the value representative for each class was tested for intra-urban mapping. Results showed that the first principal component could discriminate reasonably well between dense built-up and residential areas.

Paper Details

Date Published: 14 March 2003
PDF: 9 pages
Proc. SPIE 4886, Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II, (14 March 2003); doi: 10.1117/12.462360
Show Author Affiliations
Attilio Fanelli, Univ. degli Studi di Napoli Federico II (Italy)
Maurizio Santoro, Friedrich-Schiller-Univ. Jena (United States)


Published in SPIE Proceedings Vol. 4886:
Remote Sensing for Environmental Monitoring, GIS Applications, and Geology II
Manfred Ehlers, Editor(s)

© SPIE. Terms of Use
Back to Top