Developing an Automated Procedure for Extraction of Road Data from High-Resolution Satellite Images for Geospatial Information Systems

Abstract

There is a growing need for automated procedures that can update road databases in Geospatial Information Systems (GIS). Such procedures are in a great demand when the GIS database needs to be updated very frequently and should comprise automated means for extracting road data from satellite images and updating GIS databases using the collected data. In this paper, the requirements for developing an automated procedure which (1) extracts road data from high-resolution satellite images and (2) prepares the extracted data for use in GIS is presented. The procedure includes a region growing algorithm to obtain road features, an edge detection algorithm, an image enhancement algorithm, a vectorization algorithm, and a georeferencing algorithm. The edge detection, image enhancement and vectorization algorithms are tested using three different high-resolution satellite images and the results are presented. Their performance is measured and analyzed individually.