Urban Dynamics with Multitemporal Envisat/ers Sar Images and Multispectral Hrv Spot Optical Image Using Data Fusion Approach

In this article a method based on data fusion approach and using multitemporal ASAR-ENVISAT / PRI-ERS SAR images to analyze different behaviours (urban growth, urban decrease and urban constance) of the urban dynamics phenomenon in rain forest context is proposed. The method computes first two attributes : NDVI index used as urban / non-urban index and as prior information to improve locally the interpretation of SAR images, and the conventional fundamental SAR temporal change measure based on edges detector thanks to its great sensitivity. Then these two attributes are combined using fuzzy subsets theory. The proposed method is illustrated on images (ENVISAT’2003, ERS’1999 and SPOT’2001) of Douala urban area which has a great pressure concerning the human occupation in the urban and suburban space. The results show an important urban growth and low urban decrease phenomena. This kind of information could be useful for urban planers of this city.