Classification et évolution des tissus urbains à partir de données vectorielles

Morphological analysis of urban fabric is a relevant step in order to perform the analysis of urban dynamics and to simulate their evolutions. In this context, the objective of this paper is to test to test two relational data mining methods (Cardinalisation and Quantiles) to label urban blocks and to produce rules on their morphology. The objective is to automate the identification of urban fabrics from vector and historical database and to extract knowledge on urban morphology. Several tests applied on the Strasbourg area on historical database, are encouraging (80% of global accuracy on historical database). From the classification results, a statistical analysis of evolutions is described. MOTS-CLÉS : tissu urbain, morphologie, bases de données vectorielles, classification, arbre de décision, évolution.