Professional workers in the spatial industries are quite comfortable with the 'God's eye' view of the world. From far above, the complexities of the real world are flattened out as if they only existed in two dimensions.
Once you have lived with maps for a while, the third dimension is largely unnecessary. In one's mind's eye, mountains leap out of the contour lines, and the pitch of a roof softens the angular lines of a building.
Patterns are revealed - but at what cost? It's important to remember that not everyone is comfortable with this view of the world. For many people, a contour map is just a mess of squiggles; a house plan just a series of boxes. To see in them a landscape - or a kitchen - is a mental leap they cannot make. Famously, men don't listen, and women can't read maps.
If you ponder the question for a moment, this is quite a problem for an industry whose entire reason for existence is built on the proposition that spatial data is a better way of presenting information than, say, a spreadsheet. Maybe maps are really not as good at representing the real world as industry professionals would like to think?
Fortunately, such leaps of imagination are not quite as necessary as they once were. The science and art of visualisation is putting the missing third dimension back into mapping. How to market the fruits of the new technology is still an issue, however.
The aim of visualisation technology is to take cartographic information and give users a much better idea of what the world described by that data really looks like. At its simplest, one could argue that the earliest form of visualisation was the aerial photograph; it's a 'God's eye' view and it's still two dimensional, but at least it attempts to describe the real appearance of the land.
Pic 2: Visualisation technology has many uses. This could be valuable image for a pilot or a forester.
Mostly, however, the challenge of visualisation is for it to come off the pedestal: to show the data from low down. Given the processing power of the modern desktop computer, this is not an especially difficult proposition, but it does require that three-dimensional data be available.
Partially driven by the needs of visualisation, there is now a host of technologies for creating 3D data. Conventional aerial photography with photogrammetry will do the trick; so will laser scanning, in which three dimensions are captured directly. Radar scanning is also very popular, because of its ability to generate high-resolution datasets irrespective of the amount of daylight, or the weather.
Such techniques readily provide a 3D surface - a digital terrain model - over which an image of the same scene can be draped. This can then be manipulated in a number of ways in the computer to provide a full 3D effect. Foreground objects can obscure background objects. The resolution of objects in the field can fall off with distance, and so on.
It is also possible to take such objects into drawing or rendering programs such as are used by graphic artists so that realistic touches - clouds in the sky, sunshine on water and so on - can be added.
A lot of work is currently going into making sure that this interface is as seamless as possible. Programs such as K2VI, from Asset Information Systems, a division of Asset Forestry Limited in Auckland, appear to be typical of the latest in the breed. Not only does it allow users to visualise their data; it also allows them to interact and manipulate it directly in 3D. It allows for animation of GIS or 3D CAD data and true 3D sound within a virtual reality environment.
Conventional GIS and remote sensing suppliers have not been left out either. Leica Geosystems' US subsidiary, ERDAS, has developed an add-on module for users of its Imagine imaging software. Called VirtualGIS, it uses outputs from the photogrammetric process - orthorectified imagery - and digital terrain models to build perspective views.
Leica also offers an optional visualisation module for users of its high-end Socet Set software for digital photogrammetric workstations. Scenes from this module, or from VirtualGIS, can be linked to form animations. The user can define a series of points of view so that when the scenes are displayed one after other, the viewpoint appears to move. This is called a flythrough.
Many of these programs will run quite happily on a high-end Win/Tel desktop PC. However, there is still a strong relationship between the quality of a visualisation and the size of the computer on which it runs. High-quality visualisation requires sophisticated hardware.
Getting access to such hardware is not quite as much of a problem as it once was. In the last few years bureaus dedicated to high-performance computing have become commonplace. The Queensland state government and the University of Queensland, for instance, recently established QMI (Queensland Manufacturers' Instititute) to get products to market faster. One of QMI's first moves was to establish a virtual reality centre (called Reality Works). There are already 10 or 12 such institutions in the country, but the QMI centre is the only one that runs on a fully commercial basis.
Nevertheless, the really interesting question, at the moment, is not so much about the techniques of visualisation as about the market, and how best to address its needs.
There are already several clearly defined market niches. One important source of revenue is local government, along with the other organisations that control the built environment. All sorts of people find 3D visualisation of use here, because it allows planners to develop models that will show very realistically how a building will look in situ.
Of course architects and planners have always been able to produce 3D drawings of their buildings. The advantage of this new way of doing things is that now it is possible to place your building in among all its surrounding buildings.
This is a serious market, where clients are prepared to spend big money. Brisbane City recently commissioned Fugro Spatial Solutions and Webmap to supply Brisbane City Council with orthophotos of the full extent of the council area - 2000 square kilometres. Webmap is a joint venture between Brisbane-based Airesearch Mapping and Boustead IT of London.
The joint submission offered Webmap's existing 15 cm resolution orthophoto coverage over the urban areas of Brisbane, plus Fugro Spatial Solutions' 25 cm orthophotos of the forested areas and Moreton Bay islands.
Pic 3: A new subdivision cam be examined in detail, long before it is built
Webmap recently commissioned a similar series of high-resolution orthophotos in Sydney and Melbourne. Most of this data will be used for building digital terrain models, and from them, visualisations that will be used for planning purposes.
The concept has many advantages for planners in terms of making sure there are no surprises when the building is built. It should be possible to identify exactly whose view will be obscured by a new building, or how much of a park will be in shadow.
A second marketplace that is rapidly maturing is route planning. This was pioneered by the military and by civil aviation in flight simulators; early simulators did little more than show a line to represent the horizon. Modern versions are geographically specific - they show real landcapes and offer pilots a pretty good facsimile of the view they will really see.
This is obviously beneficial for airline pilots, who need to know something about the airports they will be landing at; it is even more so for military pilots, who can now fly a simulated mission over real terrain and identify real targets.
Much the same technology is now being developed for use in other sectors. It is possible to visualise a road or other transport corridor before it is built, so that traffic engineers can actually drive along it, checking, for instance, for features of the road that might cause accidents. Such technology can also allow people to identify places where the visual impact of the new road may be unacceptable.
The first job to come out of the QMI Reality Works centre has been the creation of a model of the proposed inner northern busway in Brisbane. The bus route has caused controversy, because it cuts through Victoria Park and passes close to a major hospital.
A third area that is opening up is in forensic analysis. Police are using 3D imaging to re-create crime scenes. Depending on the way in which one sources the data, it is possible to re-create all the significant details in the computer. Even if questions only come up later, evidence is not lost. This has the potential to be particularly valuable in investigating road accidents, for instance, where clearing up quickly after a crash is a high priority.
The uses of visualisation are diverse, but it seems there is one thing that links them: they all attempt to represent the real world, but it is a world that has been manipulated in some way. Spatial data is a very significant part of this process, but it is by no means the only part, nor even the most important.
Staff at the University of Melbourne, for instance, are about to begin trials of a new system that will present spatial data to farmers. The plan is to provide data on possible landscape futures to farmers, but presented it in a way they can understand. Visualisation technology is vitally important to this process, but there are many other inputs and outputs for farmers to consider.
Pic 4: The aim of visualisation is decision support.
By and large, the commercial justifi-cation for visualisation technology is decision support: if we spend the time and money up front to simulate environments, we will make better decisions, which will save us money and lives. Increasingly, then, we should expect to see visualisation built into decision support systems. It will be important to the designers of such systems to remember that generating a picture, no matter how pretty, is not the justification for such systems. The justification is that people will make better decisions.
Jon Fairall is the editor of GIS User
