3D buffering: A visualization tool for disaster management

Abstract

Nowadays, 2D GISs are common and their related theories, concepts and models like for geometrical modeling and spatial relationships of objects are also well addressed and investigated. Most of the tasks related to 2D GIS applications are quite straightforward and relatively easy to handle. 2D GIS spatial analysis such as proximity analysis or proximity computation, network analysis, overlay function, neighborhood function, metric measurement and other analytical operations are also well understood and well researched by the GIS community. However, problem started to surface once we move towards 3D domain, i.e. to add an additional dimension to the current 2D GIS situations such as in spatial data modeling, analysis and application. Manipulating and handling spatial objects become more complicated as we move toward 3D. This paper attempts to address one of the problems in 3D analytical operation, i.e. 3D proximity analysis. As we know that vertical component of spatial data directly interacts with the X and Y from the planimetric plane and makes the description of an object even harder to define. Data structure that supports object generation, preserves and maintains relationship with the neighboring objects is important in the 3D geometrical modeling. Several researchers in this problem domain have stated that 3D conceptual model, topological relationships, data collection, and spatial analysis might comprise a wide spectrum of questions and needs a lot of efforts to realize the solution. Although advancement in computer graphics have benefited to community in terms of 3D visualization and display but some other critical aspects like 3D spatial modeling together with the semantics information and spatial operators are hardly addressed, defined, and implemented commercially. The commercial GIS packages that able to handle 3D datasets are rather limited to surface analysis and visualizing them in 3D. GIS accepts the fact that 2.5D GIS involves a single height attached to the planimetric positions (X, Y) whereas real or true 3D GIS should able to handle data like planimetric data with multiple heights, e.g. solid objects. Some advanced 3D tasks such as 3D overlay functions, and network functions are not available in some commercial GIS software, for examples, ERDAS's Imagine VirtualGIS, Intergraph Inc's GeoMedia, PCIGeomatics's Geomatica where they provide excellent tools for 3D visualization and 3D texture models. The systems also provide some operations like surface generation, volume computation, image draping, and terrain inter-visibility can be carried. However, true 3D operations are hardly available. Inevitably, many issues need to be investigated. To move on to the 3D GIS, the third dimension must not be constrained by the single XY plane only. Considering the 3D analysis is the core component of the 3D GIS, therefore, an investigation that involves data input and 3D analytical operation will be addressed in this paper. Other aspect such as databasing is out of the scope. The developed analytical operations have been tested using real datasets that cover Universiti Teknologi Malaysia (UTM) main campus. In general, managing disaster scene is quite demanding and needs rapid spatial information on the spot where some of the required information is in the form of 3D display of a spatial query and analysis. This paper discusses the development of proximity analysis that is the 3D buffering. The corresponding algorithms that work for most of the spatial primitives, i.e. point, line, and polygon in 3D will be discussed. We tested our buffering approach by using photogrammetrically captured datasets. Finally, the paper provides outlook to the proposed work towards the development of advanced 3D analytical solutions in 3D GIS domain. © 2005 Springer-Verlag Berlin Heidelberg.