Modeling of Buildings in Urban Areas from high resolution stereo Satellite Images for Population Estimation and Change Detection

Abstract

Models of urban areas are an important input for many applications in the field of urban monitoring. Besides the creation and updating of maps from sprawling urban settlements the models are also used for simulation and planning in case of catastrophic events like flooding, tsunamis or earth quakes. With the availability of very high resolution (VHR) satellite data investigations of large urban areas regarding their three dimensional shape can be performed fast and relatively cheap in comparison to aerial photography especially for cities in developing countries. Most of the actual methods used for the generation of city models depend on a large amount of interactive work. A method for automatic derivation of - in a first step - coarse models of urban structure is therefore of great use. In this paper the methods used for such an automatic modeling is presented. The methods are based on stereo images from VHR satellite stereo imagery provided by satellite sensors like IKONOS or QuickBird. In a first step a digital surface model (DSM) is derived from the stereo data. Subsequently a digital terrain model (DTM) is generated from the DSM. Orthoimages are created in parallel using the derived DSM. Based on the local height differences between DSM and DTM and the normalized difference vegetation index (NDVI) a coarse classification of the orthoimages is performed. The classification divides the images in four classes: low level and no vegetation, low level with vegetation, higher level and no vegetation and higher level with vegetation. Based upon this classification object models are selected and object parameters are adapted to create an object-based representation of the satellite image scene - especially of man made buildings. The used processing steps are evaluated and the results for IKONOS stereo pairs of the cities of Athens and Munich are presented and discussed.