Knowledge based information mining for urban classification using multispectral high resolution images

Space imagery offers great support in various types of applications. The huge amount of information provided in a remote sensed manner facilitates the analysis of Earth surface. The image content classification is one of the first steps to follow in the data mining process. After forty years of research, this is still a topical and challenging issue. With the improving of the acquisition sensors specifications, the user requirements became more exacting. In order to give a proper solution to this problem, the present paper combines an intelligent knowledge-based information mining system with multispectral high resolution imagery. The presented application intends to provide a user-based classification for the urban area of Bucharest, Romania.