der to
enges
n turn
ation,
meet
urban
ion to
asets,
/es an
This
vithin
ed to
urban
to aid
from
an be
‘ance,
if an
years
force
and
fence
it not
iction
matic
ilding
ferent
1odel,
rated
= 3D
ed is
than
to its
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
similarity with the image seen by the human eye. Changes in
elevation are clearer as contour lines and spot heights need not be
interpreted. Intervisibility can be assessed and more detailed
surface information can be displayed, such as entrances to
buildings and height of windows. The ability to fly around
objects in real time offers the opportunity for increased situational
awareness.
However, a potential disadvantage of using a 3D model is that it
may not be realistic, surface colours and textures may have to be
inferred, vegetation, such as trees, have to be representative from
a limited library. The speed of production must be considered
since detailed 3D models can take a significant time to generate,
as well as the amount of extrapolation, such as assuming the rear
of a building to be the same as the front. As with 2D maps, issues
such as currency and ensuring the models are up-to-date still
apply. Issues of how to serve 3D and VR data need to be
explored. One option is to use web servers to allow distributed,
secure and fast access to large centrally held 3D models over a
network. New technologies are being developed that allow users
to access an image of the models without having to download
terabytes of data. The type of viewer used by the end user also
needs to be considered, with issues cost, ease of use, maintenance
and functionality to be considered.
In parallel with developments in 3D modelling and visualisation,
there have also been significant developments in sensing systems
that can rapidly collect the 3D data required to generate urban
models. Techniques such as rapid capture with terrestrial laser
scanning have allowed large datasets to be collected within a
small time-scale and this does not require excessive processing
time in order to produce models. On a larger scale, airborne
LIDAR can be used to collect data over a city at good accuracies
and resolution. These laser surveying systems are routinely used
to generate 3D models for civil engineering projects.
2. METHODOLOGY
To investigate these issues and relevant technology developments
in the creation, exploitation and visualisation of 3D urban models,
a study was recently performed on Rapid 3D Urban Modelling.
The project initially performed an assessment of the end user
community for 3D urban data and defined their information
requirements. This was completed by targeted visits to
organisations such as the Defence Geographic Centre (DGC) and
Geographic Engineering Group (GEG). Questionnaires were
used to obtain product specifications such as scale, coverage,
feature set and timeliness issues. It was important to define the
level of information required by the user, whether it be a basic
Digital Terrain Model (DTM) or a detailed 3D model and the time
frame of requirement. The visualisation level was also
considered, from full scene rendering with illumination effects or
just simplified objects. Production capability was also assessed
such as data sources, timeliness and software used.
The next task was to determine the government sponsors of work
in this field and to capture their main objectives. This was
achieved by personal communication with research co-ordinators
and evaluation of tasking orders.
A review was then undertaken of software tools that could be
used to generate, render and visualise 3D urban data, covering
both commercial software packages and academic research
developments. Information is a key part in providing 3D urban
data and therefore the next step was to provide an overview of the
available information sources that can be exploited.
3. OUTCOME OF STUDY
The study found that there were numerous users of urban data
within the defence community. Homeland defence activities are
very rapid as information is often required urgently. Presently,
Ordnance Survey (OS) data and military mapping data is used
alongside survey data from total stations. Recently, terrestrial
laser scanning was being tested for future implementation as it
offers the opportunity for rapid data capture from longer ranges, a
major advantage in anti-terror operations.
It has been suggested that 3D models of high-risk areas could be
created and stored for future use. Sites of high security value and
high risk public events in London, for example, require security
reviews, frequent contingency planning and constant updating in
the light of new requirements or changes to the environment. The
ability to walk through and view the area in different conditions
could save recurring costs of survey visits. Changes to physical
security measures such as fences, barriers, lighting and patrol
frequency can be modelled easily. The study found that there was
a need for 3D mapping although simplified models with little
rendering would be sufficient for many site orientation
applications where accuracy was not a paramount requirement.
Soldiers use 3D mapping at Brigade and Company level, with
activities such as mission rehearsal, scene familiarisation and
mission planning. Specific applications include route analysis,
location of observation posts, identification of potential obstacles,
building access, collateral damage prediction, containing and
isolating combatant forces, creating buffer zones, restoring
essential services, evacuation of groups of people and gathering
information. This group requires accurate models with high
levels of visualisation at short time scales. It would be unrealistic
to build a model of locations to support planning of urban
operations when time duration of operations cannot be foreseen
beyond a few weeks. However, for urban areas where a long time
presence is expected, or where operations might be expected to
recur over a period of time, the development of an urban model
could be a worthwhile investment for planning and to assist in
situational awareness. It was found that in practical terms,
additional information would be required with the 3D model to
achieve maximum utility. For example, building usage
(residential, commercial, military, police), construction details
(cladding, material, thickness) and significance (cultural, medical,
educational, religious) would add value to a model.
For some operations, it was found that users appreciate feature
annotation, such as heights of window sills and simple
presentation as there is not sufficient time to use complex