ABSTRACT
ABSTRACT: The automation of the extraction of road networks from high-resolution satellite imagery is a very active topic in the photogrammetry and remote sensing community. In this paper, we report on the progress towards the automated extraction of urban roads from pan-sharpened IKONOS imagery. The report focuses on highway intersection and residential road network extraction as the predominant data requirement. The authors present a semi-automated strategy to extract roads more accurately and reliably. Preliminary results extracted from pansharpened IKONOS images of an urban highway intersection scene are demonstrated. The proposed method based on the multiscale wavelet transform and the knowledge of highway geometry is developed in an interactive environment and has proven to be very promising. Aspects of both 2D and 3D highway geometry extraction are discussed. Details of a realistic error analysis are presented.
