Geographic Information System Data for Supporting Feature Extraction from High-Resolution Aerial and Satellite Images

Abstract

This article presents some theoretical and practical aspects of using geographic information system (GIS) data to support feature extraction from high-resolution aerial and satellite images. GIS data have been used to calculate parameters for objects belonging to major classes of man-made objects represented on large-scale topographic maps. A moment-based mathematical approach was used for the formal description of objects and calculation of their shape parameters. A decision tree method was employed to formulate classification rules applied to classify objects extracted from high-resolution aerial or satellite images.