Cristage: A 3D GIS with a logical crystallographic layer to enable complex analyses

Abstract

Researches on 3D GIS have mainly focused on the geometrical and topological modeling of objects to ensure data coherence and correct visualization (Molenaar, 1992; Pilouk, 1996; Zlatanova, 2000). While well suited to applications where visualization is important (urban planning, estate market.. ), these models are not suited to compute complex queries such as the propagation of a natural phenomenon (e.g. roof collapsing in a quarry). This paper presents Cristage (Poupeau, 2006), a 3D GIS prototype, that enables to perform complex analyses. Cristage is composed of two layers. The first layer deals with data coherence, visualization and computation of simple volumic queries (intersection, inclusion). It relies on Zlatanova's SSM (Zlatanova, 2000) and a spatial enumeration of objects by tetrahedrons. The second level is a logical layer on which analyzes are performed. Each geographical object is abstracted by one or several crystals and is embedded in a 3D mesh. Mesh adjacencies are stored in this level. These adjacencies are purely topological: they correspond to adjacencies between the real geographical features, and not necessarily between their geometrical representations. An adjacency is described in Cristage by a "contact object", which is a 2-Cell that bounds two 3-Cells (a 3-Cell is the topological representation of a crystal). To study geographical phenomena, we add to "contact objects" of interest a semantic information. Then a "coupling object" gathers a set of "contact objects" (topologically associated to a cell-tuple). It represents a 2D geographical object that structures 3D space. Thus, DTM, faults or stratigraphic joints are "coupling objects". The logical layer of Cristage with "coupling objects" constitutes the basis to perform spatial analysis and to assess phenomena propagation (Bonin, 2006).