Geography Reference
In-Depth Information
245
'Camera' properties
- upland landscapes such as Cumbria are invariably more impressive
in the field than when represented on a photo-like image. In part this is because the focal
length of the human eye (equating approximately to a 50 mm camera lens) has wider pe-
ripheral vision. Even ignoring the effect of depth cues through stereo vision, it is difficult
to replicate this effect when rendering computer-generated imagery. The use of the trans-
parency to augment the centre of the view with the glacial scene emphasized the impor-
tance of human peripheral vision as being a major contributing factor to the overall impres-
sion of a landscape. Different modes of interaction, such as interactive stereo widescreen
formats or more immersive techniques, may go some way to addressing these issues.
Spatial context and 'a-priori' knowledge
- the ability of observers to recognize viewpoints
in the field is affected to a degree by their familiarity with the area, through
a-priori
knowledge gained through previous visits, by studying maps beforehand or through
the spatial context gained by walking to the viewpoint, aided perhaps by a map or
positioning device. In this exercise the student groups familiarized themselves with the
area by generating the viewpoints themselves, which inevitably involves exploration of
the three-dimensional model of the area. This is significant as landscape visualization is
often used in a context where the audience is not privy to some or all of the knowledge
associated with spatial context and therefore the viewer may find it difficult to orientate
themselves when presented with a new visualization in isolation.
The 'essence' of a view
- many non-visual factors were important in contributing to the
'essence' of the view, such as weather conditions, sound and sense of exposure or the
conditions under foot. This may be significant when considering whether people orientate
and navigate in virtual environments in a similar way as they do in the real environments.
Evidence for model building
- the reconstruction of a past landscape in the form of a
retreating glacier raised an interesting issue in terms of communicating the evidence
used to create such a model. As soon as historical or hidden landscape components, such
as glaciers and geology, respectively, are added, we must put our trust in the way the
nature of evidence and processing to build such models is communicated. Despite careful
communication of the sparse and uncertain nature of the evidence to support the glacial
formation shown in Figure 12.1, many students regarded this reconstruction as 'truth'.
Clearly many of the observations above relate to the scale and complexity of the landscape
scene being modelled. Cumbria is generally regarded as a picturesque landscape and the
frames of reference for many visitors would be distant mountainous views. As soon as
the landscape scene becomes dominated by closer surface objects such as trees and build-
ings, the types of digital representation offered by the radar model become inadequate for
portraying a recognizable scene, as shown in Figure 12.2.
Replicating scenes like this requires finer resolution terrain data and a huge amount of time
and effort capturing and modelling individual surface features, often photogrammetrically.
Typically the undulations on the radar model, or higher resolution LIDAR models, would
be stripped off and CAD-like building and tree objects would be placed on the surface.
When we begin to consider what could be termed 'non-picturesque' landscapes, then scenes
dominated by surface features become more common. We should ask at this point what
exactly we should be attempting to model in a certain landscape given the magnitude of
Search WWH ::
Custom Search