Prediction of building shadows cast on urban roads based on digital elevation models and high resolution aerial imagery data

Abstract

This research presents a methodology for prediction of building shadows cast on urban roads existing on high-resolution aerial imagery. Shadow elements can be used in the modeling of contextual information, whose use has become more and more common in image analysis complex processes. The proposed methodology consists in three sequential steps. First, the building roof contours are manually extracted from an intensity image generated by the transformation of a digital elevation model (DEM) obtained from airborne laser scanning data. In similarly, the roadside contours are extracted, now from the radiometric information of the laser scanning data. Second, the roof contour polygons are projected onto the adjacent roads by using the parallel projection straight lines, whose directions are computed from the solar ephemeris, which depends on the aerial image acquisition time. Finally, parts of shadow polygons that are free from building perspective obstructions are determined, given rise to new shadow polygons. The results obtained in the experimental evaluation of the methodology showed that the method works properly, since it allowed the prediction of shadow in high-resolution imagery with high accuracy and reliability.

Keywords: Parallel Projection; Building Shadows; Solar Ephemeris; Aerial Images; Digital Elevation Models; Airborne Laser Scanning Systems.