The use of high resolution digital surface models for change detection and viewshed analysis in the urban area around the pyramids of Giza, Egypt

One of the biggest threats to cultural heritage is related to their rapidly changing and developing surroundings. The Giza pyramid plateau is a prime example of this phenomenon, as it is threatened by the enormous urban expansion of Cairo over the last decades. Documenting, monitoring and modelling such a pressure requires accurate and detailed geographic data, which can be derived from recent up-to-date, high resolution satellite images. Remote sensing techniques have proven to be very useful to visualize and analyze urban sprawl and land use changes in two dimensions. The impact assessment of urban sprawl near specific heritage sites, however; needs to be complemented with accurate 2.5D-information. In an attempt to do so, digital surface models (DSMs) from Ikonos-2 (2005) and GeoEye-1 stereoscopic images (2009 and 2011) have been computed in order to analyze recent urban changes. Change detection methods are mainly developed for large scale high resolution aerial images; however this paper focuses on the one hand DSM creation and its challenges resulting in an improvement of 2.5D change detection method for small scale satellite imagery in mainly informal areas. On the other hand a view shed evolution is presented. The combination of the enhanced digital terrain extraction (eATE) module of Erdas Imagine^® and ground control points collected in the field provides accurate and high resolution DSMs. The impact of shadow and different urban morphologies however influence the pixel-wise comparison of the two DSMs, which results in different approaches for different city districts. The resulting 2.5D change model clarifies not only the urban sprawl, but also the increase in building levels, directly related to pressure on the famous pyramids. This pressure is furthermore analyzed by creating different view sheds through time from the plateau towards the city and vice versa. An integration of population statistics complements the model, hence allowing- it to become a useful policy instrument.