Monitoring the Urban Expansion of Cairo from 2004 to 2010 through SAR Data using a Non-Parametric Supervised Classifier
Delgado Blasco, J. Manuel1; Verstraeten, Gert2; Hanssen, Ramon3
1Delft University of Technology & KU Leuven, BELGIUM; 2KU Leuven, BELGIUM; 3Delft University of Technology, NETHERLANDS
In this study the urban expansion of Greater Cairo from 2004 to 2010 is monitored. Synthetic aperture radar data was selected based on availability, coverage and weather conditions, producing several interferometric coherence products suitable for the multi-temporal analysis of urban coverage. Imagery from two different tracks of SAR data of Envisat ASAR sensor were selected. In total 21 coherence maps satisfying the conditions of low perpendicular baseline and low temporal baseline. Data acquired during March-April were removed from further analysis as the interferometric coherence was seriously affected by sand storms that are typical in this time period. Hence, 9 coherence maps and 38 calibrated SAR images were used as input for subsequent classification of urban and non-urban areas. In addition, several designs of neural networks combining different inputs, number of perceptrons and training algorithms were tested for selecting the best performing design. Finally, classification maps 2005, 2009 and 2010 for the two different tracks were produced, allowing to evaluate the evolution of the urban expansion of greater Cairo. Two trends were identified: a densification in the cities and villages inside the floodplain, and a sistematic urban expansion on the eastern and western plateau.