Three-Dimensional GIS

Need for 3D

Current GIS are generally limited to two horizontal dimensions, which disregards the third dimension (height or elevation). The following example illustrates the necessity for considering information in the vertical dimension. Figure 1 illustrates how noise emission caused by a railway would be calculated by current systems. Applying the usual two-dimensional data analysis tools leads to the conclusion that none of the buildings appears to be impacted by noise emission from the railway system.

Figure 2 shows the same example but not from the bird's-eye view. When looking at the scene from the side view, it is clear how traditional analysis leads to an incorrect result. Since the land surface has not been considered in 2D analysis, both buildings seem to be protected against noise by the noise barrier. In reality, however, the right building is affected by noise [Page 471]emission since it is located higher and thus above the sound barrier. To calculate these impacts of the terrain, 3D analysis is required.

Figure 1 2D Data Analysis

Noise emission calculated by a traditional 2D GIS.

Previous Ways of considering the Third Dimension

Due to technical limitations, full consideration of the third dimension has not yet been fulfilled by commercial GIS. Nevertheless, since it is clear that the third dimension must be considered, several approaches have been developed to get and use height information in spatial data processing.

The most widely used method involves the use of a raster of height values in which each cell value represents an estimate of the elevation of the surface at the location of the cell. Such so-called 2.5D digital elevation models (DEM) can be used to visualize 3D surfaces in a number of ways, including using simple color coding of elevation ranges or the calculation of illumination shading based on a derived slope surface. Furthermore, many systems enable their users to calculate and draw contours. Since this kind of visualization of spatial information is similar to that used on traditional paper maps, many applications requiring the representation of surfaces can be satisfied by this method.

Although 2.5D GIS answers some of the needs for 3D representations of spatial data, it cannot provide a solution that is close to reality. Raster DEMs represent space as if the real-world relief consisted of cubes and thus does not represent the world as a continuous surface (see Figure 3). Also, such surface representations do not provide a true 3D representation since any single (x, y) coordinate pair can have only one z-coordinate. True 3D representations would allow more than one point to exist at any (x, y) location.

...