GIS – 3D and Beyond
Alias Abdul Rahman
Department of Geoinformatics
Faculty of Geoinformation Science and Engineering
Universiti Teknologi Malaysia
81310 UTM Skudai, Johor
Spatial information system or GIS deals with spatial data collection, processing, manipulation, databasing, and visualization as well as the derived information. GIS evolves from spatial data of 2D to 2.5D data type. Today, more and more advanced applications requires more functionality from such spatial system, e.g. applications in advanced urban planning, geology, oil and minerals exploration, and underground mapping including utility mapping. Present commercial systems are hardly able to offer solutions for advanced analysis like 3D spatial analysis. This paper describes some efforts towards the realization of 3D GIS and what needs to be done for next generation GIS. The paper also highlights some of the issues and problems associated in developing such advanced GIS system. The findings of research in this problem domain reveal an exciting new paradigm for next generation of GIS software and system, i.e. beyond 3D domain.
Until today, GIS remain one of the exciting tools to manipulate geospatial data. Systems offered by various vendors are growing in terms of functionalities and all of these systems are basically two-dimensional (2D) based and at the most just able to manipulate another data layer like contours or heights to the existing datasets or layers. Having this system around means we have 2.5D based GIS. Now, GIS users are getting more complex datasets and need to manipulate these datasets and generate information as we perceived in the real world, i.e. in 3D environment. This environment provides much better understanding of the geospatial pattern and phenomena, either in small or large scale areas. The aim of this paper is to examine the current GIS software developments and some research efforts towards realizing the new breed of GIS systems or software that able to provide 3D spatial information. This paper also attempts to look further on the possibility of extending the current spatial data modeling techniques to 4D and beyond, i.e. n-dimensional (nD).
Section 2 looks into the current situations or in other words the 2D GIS. Section 3 describes the 2.5D GIS while the 3D GIS will be discussed in Section 4. Section 5 discusses ubiquitous GIS. Finally, the conclusion and some further works in Section 6.
2. TWO-DIMENSIONAL (2D) GIS
In general, most of the spatial data modeling engine of any GIS software is based on spatial primitives of points, lines, and surfaces or polygons. The modeling technique is quite straight forward where the notation, convention and the formalism mechanisms are fully understood by many GIS users and software developers. It is not the intention of this paper to describe those 2D modeling formalisms since they have been solved and fully developed. Tremendous efforts have been done and the 2D systems are being utilized and implemented in various GIS applications. The systems are being well received by the GIS community as a whole.
3. TWO AND A HALF-DIMENSIONAL (2.5D) GIS
This is an extension of the system discussed in the preceding section that is by adding height to the existing datasets such as height data from contours, heights from photogrammetric workstation or any other acquisition techniques like GPS or Lidar data acquisition system. At the moment some GIS software offers surface data manipulation module as part of the typical GIS modules or programs. The module is just to manipulate heights and do some surface analysis like contouring, slope and aspect and other computations on top of the typical 2D data layers. The author considers 2.5D GIS is the starting point that one should consider in developing such 3D system. The modeling component is very interesting to look at. Many research efforts have been done in this spatial data modeling domain such as Molenaar (1991), Pilouk (1996), Zlatanova, and Peng (1997). Molenaar proposed a Formal Data Structure (FDS) to link between various primitives as illustrated in Figure 1. The basic relationships of object primitives is quite clear and could be utilized to certain extent but not quite suitable for GIS data with heights as we have in digital terrain model (DTM). The model then has been modified by Pilouk (1996) for a spatial system as we called it a 2.5D GIS. As a result, several GIS software or systems are based on this model. Manipulation of terrain data together with other features is now possible by using the 2.5D data model. Query like “show a group of polygon features on the terrain” could be performed quite nicely (see Figure 2).
Figure 2. Queries from 2.5D data model (Courtesy of Pilouk)
4. THREE-DIMENSIONAL (3D) GIS
It is interesting to note that the Molenaar’s data model could be extended for 3D GIS software development. Several researchers like Zlatanova (2000) and Abdul-Rahman (2000) have utilized the model for manipulating 3D spatial objects in their works. The adapted model as illustrated in Figure 3 shows the relationships of 3D object’s primitives like nodes, lines, surfaces, and solids. The model works for both data in raster as well as for vector datasets.
Figure 3. The 3DFDS model
Abdul-Rahman (2000) manipulated 3D objects via 3D raster-based (voxel) objects approach and had generated information via proprietary object-oriented DBMS (i.e., POET DBMS –Persistent Object and Extended Technology). The results show the 3DFDS is capable of manipulating 3D spatial objects as the objects were created via 3D triangular irregular network approach.
Figure 4. The 3DTIN and 3D raster polygons (in slices)
Zlatanova (2000) implemented the modified version of 3DFDS for Web-based 3DGIS. The model works in Oracle spatial DBMS as the 3D objects were constructed in VRML coding environment, see Figure 5.
Figure 5. The Web based query on the generated 3D objects (Courtesy of Zlatanova)
The development of 3D GIS is growing and many works are being done in several research centers and universities as indicated by Abdul-Rahman (2006).
5. BEYOND 3D GIS
The author believes that one day GIS will have a true 3DGIS in near future. This is based on the current pace of research efforts that being done in various centers and universities in some parts of the globe. Current trends clearly show that GIS users demand more than the current technology could offer. Here, future users would like to have information of a particular object in a certain regions or areas in a split of seconds, very accurate, and easy to access either standalone or Web/Internet solution (Hunter and Tao, 2002). Although theoretically, the spatial modeling of objects could be extended to multi-dimensional (nD), the computing visualization systems only permit up to 3D environment. This section attempts to highlight some possible research works that could lead to the “future” GIS or ubiquitous GIS. The author believe that this is the future trend of GIS where every component in GIS like data collection, data manipulation, databasing, and reporting (and visualization) were done seamlessly and they are highly dependent on mobile computing environment. We have seen several research groups are working on this direction e.g. GeoICT Lab, York University, Toronto, Canada; GIS Section at TU Delft, The Netherlands, and Fraunhofer Institute (IGD), Darmstadt, Germany. This newly established research initiative in GIS focuses on components integration and mobile, thus the system could end up in small size with some intelligence-built in components. The following figure shows basic configuration for the ubiquitous geospatial system. In general, voice sensor, small display, small power unit, precise location finder (like GPS or Galileo), wireless communication system, and equipped with compact size computer could be parts of the whole system.
Figure 6. The ubiquitous GIS system (Courtesy of Tao)
This new mobile system requires more research efforts to address some issues. It has been recognized that the following would contribute to the development of such system.
6. RESEARCH EFFORTS TOWARDS 3D GIS AT UTM SKUDAI
Efforts to realize true 3D GIS are taking place. We address some of the issues like 3D spatial data modeling, 3D visualization navigation, and 3D databasing. Below is the summary of our current PhD research works:
6.1 3D topology for 3D GIS
This research attempts to produce a new topological model that works for 3D spatial objects, and eventually new breed of 3D spatial analysis. New 3D spatial operators are anticipated from this research project.
6.2 Spatial data modeling for 3D cadastre
Since we want to explore the possibility of providing sound information to the land and property owners then the project has been formulated and geared toward the direction. This is inline of the government policy – that is to provide fast and accurate service to the community.
6.3 3D navigation for 3D GIS
This is to enhance the current topological research work. Navigation will provide means to locate and explore the 3D spatial objects in great detail interactively. The project incorporates some aspects of VR and dynamic 3D visualization.
6.4 Managing 3D spatial objects for large area
It is our aim to have a 3D spatial database that caters for large areas, i.e. statewide or nationwide. National mapping agency would be one of the possible research targets and users.
6.5 3D utility mapping and databasing
At the moment most of our spatial plans are based on surface based features. However, it is quite hard to have a good system that able to manage subsurface utility objects like pipelines, power cables, telecommunication cables including pay-TV cables, gas, water pipelines, and other utility lines. These objects require 3D utility mapping and 3D database.
7. SUMMARY AND OUTLOOK
The paper has described the evolution of GIS, i.e. from traditional GIS systems to ubiquitous version of GIS. The real concerned of this paper is the spatial data modeling, it is the core component of any GIS software development. This is the core component of any GIS software development. Further research is required to investigate methods of spatial data modeling in mobile environment especially for 3D spatial objects so that sound 3D geo information could be generated and produced in situ. Significant testing on the developed model like the CoS (Condensed Spatial) spatial model is also required to ensure that all the object primitives are fully interrelated. Lastly, with regards to ubiquitous GIS further investigation with respect to interfacing to the existing systems are also required.