Bégué Agnès, Borne Frédéric.
1999. Scale-dependent approaches in land-use cartograhy using remote-sensing data.
In : Scaling methodologies in ecoregional approaches for natural resources management. Kam Suan Pheng (ed.), Chu Thai Hoanh (ed.). IRRI
Autre titre : Approches liées à l'échelle dans la cartographie d'utilisation des terres à partir de données de télédétection
Résumé : Two scale-dependent approaches in cartography are illustrated using remotely sensed data. The first example deals with a "bottom-up" method. PAPRI is an original segmentation method based on image local texture analysis. This segmentation requires the interactive definition of training areas over any type of satellite product (gray-level raw images or classified images). The result is a segmentation of the image in homogeneous parts of space (what is called here "landscape"). Examples of "landscape" mapping using raw radar images and classified SPOT images are presented. The second example covers "top-down" methods applied to assess land cover percentage using low-resolution images (NOAA-AVHRR type). Coarse-resolution sensors are appropriate for work at a regional scale at reduced costs, but the size of the pixels may not be acceptable for cartographic applications. To overcome this limitation, we compare and discuss two methods in the context of assessing the forest cover proportion of tropical areas: the TREES method based on spatial modeling and the COCKTAIL method based on spectral mixture modeling.
Mots-clés Agrovoc : cartographie, télédétection, utilisation des terres, imagerie, modèle mathématique, forêt
Mots-clés géographiques Agrovoc : Madagascar, Brésil, Pérou, Cameroun
Classification Agris : U30 - Méthodes de recherche
Auteurs et affiliations
- Bégué Agnès, CIRAD-AMIS-AGRONOMIE (FRA)
-
Borne Frédéric, CIRAD-AMIS-AMAP (FRA)
ORCID: 0000-0003-3251-8253
Autres liens de la publication
Source : Cirad - Agritrop (https://agritrop.cirad.fr/391930/)
[ Page générée et mise en cache le 2024-04-26 ]
