Community Research and Development Information Service - CORDIS

Abstract

This paper describes a procedure being developed to generalise satellite image data products automatically into cartographically acceptable land cover maps. The method initially involves low level generalisation procedures concerned with smoothing a detailed neural network classification to form primitive spatial units, followed by a higher level problem-solving stage. In the higher level stage an expert system approach is used to optimise the decision-making process in macroscopic generalisation - that is, to decide how to group primitive spatial units of given spectral classes into larger map regions matching a pre-specified land cover nomenclature.

Additional information

Authors: GOFFREDO S, JRC Ispra (IT);WILKINSON G G, JRC Ispra (IT);FISHER P F, University of Leicester, Department of Geography (GB)
Bibliographic Reference: Paper presented: The Remote Sensing Society's Annual Conference, Chester (GB), September 16-17, 1993
Availability: Available from (1) as Paper EN 37793 ORA
Record Number: 199311273 / Last updated on: 1994-11-29
Category: PUBLICATION
Original language: en
Available languages: en