Community Research and Development Information Service - CORDIS


This paper presents an original automatic processing chain for the spatial and thematic generalization of satellite-derived maps to be introduced in a Geographical Information System (GIS) database. The prototype system uses image understanding and artificial intelligence techniques, combined with fundamental cartographic generalization principles. It works individually on objects, closed polygons detected by Edge Detection technique from the original satellite image. The input to be generalized consists of an Artificial Neural Network classification of the original satellite image. The output consists of a map of main homogeneous patches, each of which is individually identified by a unique CORINE land cover class.

Additional information

Authors: GOFFREDO S, JRC Ispra (IT)
Bibliographic Reference: Paper presented: 17th International Cartographic Conference ICA/ACI, Barcelona (ES), September 3-9, 1995
Availability: Available from (1) as Paper EN 39251 ORA
Record Number: 199610096 / Last updated on: 1996-02-16
Original language: en
Available languages: en