Raster to Vector in 2D Urban Data

The aim of this work is to propose a refined approach to urban object recognition and extraction exploiting a priori information about geometrical features of the urban objects. In particular, the proposed procedure shows that it is possible to improve the characterization of elements of the urban scene by assuming that they fit to some geometrical models. Advantages includes the possibility to manage urban object as vector files, to compare object presence and shapes in multi- temporal remotely sensed data sets, and finally to compare and fuse remotely sensed data with Geographic Information vector layers.