Flood monitoring in urban areas: statistical vs. neurofuzzy approach

Abstract

Absfracf-The paper aims at investigating different classification and segmentation tools for flood monitoring using satellite S A R images. To this aim, two different approaches, namely statistical segmentation and neurofnzzy classification and compared and discussed. The methods show, in general, the possibility to provide to a good extent accurate maps of the flooded areas using simple processing schemes. This stresses the effectiveness of satellite S A R images for real-time flood monitoring. T. M a d Peliizzen, P. Lombard0 and A. Toaora are INFOCOM Ept. -University of Rome"La Sapienza", Via Eudossiana 18 ~ 00184 Rome, Italy University of via ~~~~t n I, 27100 pavia h i y (e.maii: All the images are in PRI or similar format (12.5 meter (e-mail: t.macn.pelli-n@infoeom.uniromal.it). P. Gamba and F. DelI'Ayua ace %th Depaameot of Electrooicsf.dell 'acqua@ele.unipv.il).

A collection of data for urban area characterization

“…The 2000 image is also the nearest in time available image of the area after the last flood of the river Ticino, which flows through the Southern part of the town. A few algorithms have been designed to exploit this set and optimize change detection in SAR data [11] or change detection using multi-source data, from SAR and multispectral sensors [12], [13].…”