Georgopoulos, Andreas 
  
The visualisation output, in most of the cases, may combine plans, vertical views or oblique perspective views on plain 
paper or on a display and even more virtual walk-through or fly-through in the portrayed area on a computer or video 
device. 
The aim of this paper is to describe the methodology for producing the three-dimensional representation of traditional 
settlements. The major steps to this end are the data processing, the three-dimensional modelling and the adopted 
visualisation method. All the above are evaluated for their applicability and interoperability. 
As an example, the three-dimensional visualisation of a traditional North Aegean settlement, which was destroyed by an 
earthquake about 30 years ago, is presented. The available data were small scale aerial photography before the natural 
disaster, old non-metric oblique photographs of the settlement and an up-to-date detailed survey of the terrain. Using 
these raw data and applying suitable methodology the three-dimensional visualisation in large scale of the area, as it 
was before the earthquake, was produced. The vector result was then rendered to make the final visualisation product 
more realistic. Such an effort produces a result, which is considered as invaluable in terms of restoration and revival of 
a destroyed settlement. 
2 DESCRIPTION OF METHODOLOGY 
Considering the needs of a large scale representation of a small settlement, it should include, in general at least, the 
portrayal of the earth’s surface and the settlement’s buildings. Moreover, the representation should include the 
appropriate level of physical or technical details, in order to achieve a considerably realistic visualisation output. Thus, 
the earth’s surface should be enriched by portraying existing streams, roads, or other relevant details and the buildings 
should be presented with their roofs, doors, windows, balconies etc. It is of significant importance to point out that there 
is no need to represent each individual building exactly as it was before the earthquake at this scale of visualisation. 
What is really important is the overall image of the settlement, while at the same time its architectural character is 
represented as accurate as possible. The above concept led to the decision of adopting a visualisation model based on 
information acquired from existing amateur photographs. The visualising output was based on a 3-D wireframe model 
of the settlement, which was derived by analysing the elementary construction forms of the buildings, while preserving 
their architectural character, instead of a more complex model with attached raster images (Nakos and Tzelepis, 1998) 
derived from the amateur photographs. 
RECENT OLD AERIAL AMATEUR 
SURVEY PHOTOGRAPHS PHOTOGRAPHS 
+ 
   
    
  
  
  
  
  
  
  
  
RENDERING VISUALISATION 
  
  
Figure 1. The outline of visualisation methodology 
The topography of the area under study can be digitally described by producing a Digital Terrain Model from data 
derived from recent large scale surveys. The TIN method gives more reliability to large scale representations and the 
ability to include physical or technical details (i.e. streams, roads etc.) in the model as breaklines. Therefore the TIN 
method was utilised for the project. Of course, the locational information of the existing details was not present in the 
recent survey and thus it was necessary to extract it from aerial photographs taken before the earthquake. The shape of 
  
286 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B5. Amsterdam 2000.