formation of the model in a certain coordinate system. Since all 
of these necessary points were not necessarily measured during 
the field work of the geometric documentation, it was decided 
to use the available data in order to virtually determine them. 
The available data, for this case, consisted of a list of 
coordinates of control points in the geodetic system and various 
vector drawings containing the desired edges and the control 
points. The coordinates of the latter were, however in different 
coordinate systems, as every projection plane carried its own 
system for convenience reasons (Daskalopoulos 2002, Demiridi 
2003). 
     
Whi ^ 
Figure 3: Geodetic and photogrammetric point cloud 
Firstly all geodetically measured control points were imported 
into the 3D model's unique drawing, with the help of a simple 
VBA routine, which was developed. Then, the exterior and 
interior surfaces of the church were divided into a number of 
sub-surfaces, defined by suitable groups of vertices. All chosen 
sub-surfaces were, within reason, assumed parallel to the 
respective projection planes of the drawings. This assumption 
meant that the positioning of the vertices relative to the control 
points was approximately correct taking into account the 
accuracy of the rectification method. 
  
Figure 4: The wireframe model 
This methodology was based on the capability to change the 
user coordinate system (UCS) by defining three points (origin, 
x-axis, y-axis) within AutoCAD”. Working separately on each 
sub-surface, the same UCS was selected by using the same three 
control points every time, both in the environment of the 
drawings and in the three dimensional model. Since the 
coordinate systems were the same, all that was needed was to 
copy the vertices from the 2D drawing and paste them into the 
3D model (Figure 4). 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
3.2 Surface model 
A major drawback of this methodology was the fact that certain 
parts of the church, such as the dome, could not be included into 
the wireframe model, because the critical vertices could not be 
determined, as the sub-surfaces were not planar and hence could 
not be considered parallel to any projection plane. Surface 
models may act as a basis for the creation of the textured model, 
when the proper surface layering is utilized. Therefore the next 
step was to transform the wireframe into a surface model. All 
the surfaces were defined and created separately within 
AutoCAD, in three different ways, depending on the nature, the 
shape and the properties of surface. This was carried out with 
the help of the “pface”, “edge surface” and “region” commands 
of the software. 
In order to complete the model of the exterior of the church, the 
dome and parts of the roof which were missing had to be 
separately reconstructed. The surface of the north hemisphere 
resulted through a surface simulation of a 3D vector drawing of 
the dome, which was the result of a digital photogrammetric 
restitution from aerial large scale pictures taken from a 
helicopter. The same 3D vector drawing was utilized for 
defining the missing vertices in order to complete the roof 
section of the model (Figures 5 and 6). 
  
Figure 6: The surface model of the interior 
3.3 Textured model 
The data used for the formation of the wireframe and surface 
models were all vector drawings of the project of the geometric 
documentation, the coordinates of the control points and all 
  
     
    
   
    
   
   
   
    
    
    
    
  
  
  
  
  
  
  
     
   
   
  
  
  
  
  
  
    
   
   
  
  
  
  
  
    
    
pr 
Al 
for 
of 
wi 
Wa 
thi 
me 
  
Fc 
m 
er 
re