Abstract
The photogrammetric method of DTM generation in built-up areas is investigated. New filtering and classification approaches are presented.
Results of practical tests with the suggested processing of the 3D point clouds are analyzed. The obtained results of two tests (16 cm and 17 cm in the standard deviation and 14 cm and 18 cm in the 68% quantile) using large format images from 1000 m altitude are considerably better than the results with the raw data.
The potential of the photogrammetric method for DTM generation is summarized and a comparison with airborne laser scanning is made. Applications like the extraction of buildings from dense point clouds or the completion and updating of existing DTMs could take advantage of using the photogrammetric approach.