Segmentation of urban areas in spot images using MRF

Classification of urban areas from a density measure is required in many application fields. For instance, propagation models for cellular radio networks are parameterized by coefficients whose value is linked to the nature of the soil. Geomarketing is also interested in knowing the population concentration (urban, suburban ...). This paper presents the first results about the construction of an algorithm devoted to the classification of urban areas in SPOT panchromatic images. This algorithm has been inspired by [5, 3] and we point out the differences between their method and ours.