Flood hazard assessment using panchromatic satellite imagery

Ahmed Shaker, W. Y. Yan, Man Sing Wong, Nagwa El-Ashmawy, Bahaaeddin I.Z. Alhaddad

Research output: Journal article publicationConference articleAcademic researchpeer-review

4 Citations (Scopus)

Abstract

Panchromatic (PAN) satellite imagery has been used successfully in different applications such as topographic mapping and terrain modelling. Nevertheless, PAN imagery is still one of the less used digital sources for land-change studies except few processes where the high-resolution of the PAN images is used to improve the visualization quality of the multi-spectral images. This research aimed to study and facilitate the use of the PAN satellite imagery in image classification for flood hazard assessment application. In particular, this research is investigating the potential use of PAN satellite imagery for flood hazard assessment of one of the high floods of the Nile River occurred in Year 1998. Several existing techniques and approaches used for digital image processing were examined and assessed for PAN image classification. The study focuses on four different approaches that could be used for PAN image classification and flood hazard assessment: a) image interpretation, b) edge detection, c) pixel-based image classification, and d) texture analysis (TEX). The described approaches were investigated for PAN satellite imagery covering a part of the Nile Valley in Egypt. Two SPOT PAN satellite images covering part of the Nile Valley in Egypt before and after the 1997/1998 Nile flood have been utilized. Different areas of interest (islands, coastal areas, etc.) have been identified to study the efficiency of the previous classification approaches for PAN image classification. The results of four PAN image classification approaches are presented and assessed for the study sites using area and sample comparative analysis. The results revealed that Contextual Classifier on PAN imagery and Maximum Likelihood Classifier on TEX imagery offer the nearest estimation of flooding areas. The study shows high integrity of the tested approaches for PAN image classification and accuracy comparable to conventional signature-based classification technique of multi-spectral images could generally be achieved.
Original languageEnglish
Pages (from-to)881-886
Number of pages6
JournalInternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences - ISPRS Archives
Volume37
Publication statusPublished - 1 Jan 2008
Event21st Congress of the International Society for Photogrammetry and Remote Sensing, ISPRS 2008 - Beijing, China
Duration: 3 Jul 200811 Jul 2008

Keywords

  • Classification
  • Flood
  • Hazard Assessment
  • Panchromatic
  • Texture Analysis

ASJC Scopus subject areas

  • Information Systems
  • Geography, Planning and Development

Fingerprint

Dive into the research topics of 'Flood hazard assessment using panchromatic satellite imagery'. Together they form a unique fingerprint.

Cite this