A Projective approach to handle 3D spatial data

1. Introduction The choice of relevant data models and data structures is still an important issue in 3D GIS. In the 3D urban GIS context, boundary representation models are usually adopted with different levels of " topology " integration In this work, a new approach based on projective properties of spatial objects has been adopted. The theoretical framework behind this is the Dimensional Model (DM), whose originality is to be defined in a projective space instead of a topological space. This model proposes a new partition of the spatial objects, called dimensional elements, and a new type of spatial relationships, the dimensional relationships. The aim of the study is to define spatial data models, and after that data structures, based on the DM theoretical framework, allowing the extraction of the dimensional relationships without point coordinate computations. The results are three boundary representation models called Spatial Dimensional Data Models (SDDM), which present interesting properties for 3D modelling.