Abstract | Full Paper | PDF | Printer friendly format

Page 1 of 6
| Next |

1. Introduction
The combination of Virtual Reality (VR) and GIS is not really a new topic. From the early 1990s, scientists began to explore the possibility of using 3D visualization technologies to interactively represent urban buildings [Day 1994]. One of the major tools they used is VRML, a high level objected-oriented language for the description of scenes and behaviors of 3D objects. VRML has been widely accepted as an international standard of interactive visualization since 1997, when ISO VRML97, or so called VRML2.0, is released [Web 1].

Kate Moore [Moore et al. 1997] started the project of The Virtual Field Course which used a VRML interface to explore a geographic database and multimedia resources fieldwork areas. The problems of geo-referenced VRML, accessing position by VMRL sensors and changing the theme of information had been addressed. Year 1998 saw more expansion of VRML interfaced GIS. Shan [1998] integrated CAD/CAM, DPS and GIS data in a desktop environment based on the 3D Web. Terrain, buildings and their appearance have been successfully modeled and rendered. Coors and Jung [1998] created GOOVI-3D, a prototype system that provides access and interaction with a 3D spatial database over the WWW. GOOVI-3D was implemented in Java and VRML using Common Object Request Broker Architecture (CORBA) for accessing the data warehouse. In their work, the flexibility of VRML was explored by proposing two lightweight extensions of VRML: an integrated name dictionary management and a SQL node. At the same year, several Korean researchers [Lee et al. 1998] proposed a Web 3D GIS with spatial analysis functionality, featuring design of the module of Spatial Operation Manager with operations such as Near Analysis, 3D Buffering, Distance Measurement and Lantern Selection. In the last three years, researchers tend to develop more specific geo-related Web 3D applications in the area of architectural and archaeological restoration [Caiani et al. 2001], meteorological service [Chan et al. 2001] and ocean science [Campbell et al. 2002, McCann 2002].

VRML has been the favorite of most Web 3D GIS researchers for over ten years in that it is cheap, platform-dependent, can provide middle-quality interactive visualization and has high compatibility with other tools like Java.

However in most work mentioned above, VRML is mainly used as a display tool, especially in the early server-side Web GIS system. In a server-side Web GIS system, such as a CGI-based one, after 2 queries from users reach the Web server, the server will analyze the queries, retrieve relevant data from the data server and generate a new VRML file for displaying on the client side, i.e. the Web browser. Users mainly have to rely on a 2D web interface to submit their queries. The potential of VRML as interactive visualization tool has not been fully employed yet. The later research work has shown a tendency of choosing a client-side structure which passes the task of generating VRML file to the Web browser itself. For example, a Java applet or a ActiveX plug-in. However, for the visualization of a large-scale area, generating VRML nodes on the air will be too expensive in terms of interaction speed.

More work is needed to explore the possibility and capability of VRML to visualize geo-information covering a large-scale area at the client-side. The best case would be 3D terrain visualization on the Web. Terrains often work as backgrounds for a lot of other information such as urban buildings, road topology, attributes distribution, and so on. A seamless combination of terrain visualization with other information layers is crucial to a good Web 3D GIS.

In this paper, we propose a selective visualization method for 3D terrain on the Web. The prototype implementation is integrated with a hybrid 2D-3D user interface design. The 3D terrain model we used is constructed from a 2D contour data file of the Singapore terrain

Page 1 of 6
| Next |