par Frauman, Evelyne ;Wolff, Eléonore
Référence Proceedings of SPIE - The International Society for Optical Engineering, 5983, 59830A
Publication Publié, 2005
Article révisé par les pairs
Résumé : In this paper we examine the possibilities of using VHR images for locating major changes (i.e. buildings, roads and vegetation) in order to focus ground topographical measurements and to keep a topographic 1:1000 database (UrbIS) of urban areas up-to-date in between two aerial photographic flights. The techniques implemented are multi-temporal classification and post-classification change detection. These techniques are used as such and in GIS-based change detection. This latest is implemented with masks exported from information contained in the UrbIS database. Main changes are well detected, though a number of "false changes" is detected and remain significant. Solutions to reduce the number of false alarms will be investigated through the application of post-classification rules.