GEOSCAN, résultats de la recherche


TitreA semi-automated approach for the recognition and extraction of water features from Landsat 7 imagery in northern Canada
AuteurArmenakis, C
SourceNext generation geospatial information: from digital image analysis to spatio-temporal databases; par Agouris, P (éd.); Croitoru, A (éd.); ISPRS Book Series vol. 3, 2005 p. 59-67
LiensOnline - En ligne
Séries alt.Secteur des sciences de la Terre, Contribution externe 2004246
ÉditeurA.A. Balkema Publishers |a Leiden, The Netherlands (Leiden, The Netherlands)
Documentpublication en série
Mediapapier; numérique; en ligne
Lat/Long OENS -66.0000 -65.5000 67.7500 67.5000
Sujetseaux de surface; analyses de l'eau; prospection de l'eau; imagerie par satellite; images du satellite LANDSAT; lacs; rivières; analyses spectrographiques; analyse spectrographique; analyses spectrales; satellite LANDSAT; images du satellite LANDSAT; hydrogéologie; divers
IllustrationsLandsat images; tables; equations
ProgrammeLa géomatique à l'appui du développement du Nord
Résumé(disponible en anglais seulement)
A common approach to automatically extract water bodies from multi-spectral imagery is by using land cover supervised or unsupervised classification. However, the exis-tence of other features with similar reflectance, leads to low separability of the thematic classes and increases the confusion level of the results. This results in low accuracy classification and low reliability requiring afterwards significant amounts of interactive editing. A semi-automated approach for the extraction of water features from Landsat 7 ETM+ imagery is presented in this work based on image processing and spatial analysis tools, combined with the spatial constraint of the terrain slope. The implementation of the methodology was evaluated based on a feature-based change detection approach between the “true” and the newly extracted data sets. The changes, commission and omission errors, are defined as the non-intersection of two polygonal data sets of identical themes. The proposed approach is promising as it detected and extracted successfully a significantly large percentage of the water body areas. Further tests are required to determine if the proposed methodology in this study can be generalized and transferred to op-erational environments.