projection. One must however still consider the typical 
limits of architectural photogrammetry, in which sudden 
variations in the object’s depth are the cause of the fact 
that small variations in camera position can substantially 
change the extent of the object’s surface, which is visible in 
the photograph. 
The photographs of the interior have a mean scale of 
about 1:150, because they had to allow for plotting at 1:25. 
One of the two photographs has been digitizedby a 
scanner of 2000 dpi (dots per inch) and with a range of 
256 grey values, using a CCD camera mounted on the 
Digicart 40. The size of the corresponding resulting .TIF 
file resulted to be about 6 Mbyte. 
In order to carry out a qualitative comparison of the 
results, the other photograph has been read at 256 grey 
level resolution with a commercial scanner of 300 dpi, 
generating a .TIF file of about 1 Mb. 
The two photographs used for the realisation of the 
rectifications have first been printed in black and white on 
a 18x18 cm^ format. Successively, the obtained print has 
been digitized by a scanner of resolution 300 dpi and 256 
grey levels, generating two .TIF files of about 4 Mb each. 
The scanning requires a few dozens of seconds for the 300 
dpi resolution. All things considered, the data transfer and 
change of format prior to the actual processing require a 
much longer time. Already in this situation, it becomes 
evident that one of the practical problems of such a 
procedure are the hardware conditions and the 
characteristics of the network. 
It is also worth noting the characteristics of a thus 
generated images. The representation of both on a 
videoscreen with a Matrox board driven by a Compaq PC 
486 of 50 MHz, does not show large qualitative differences 
between the various images. When the analysis is repeated 
after having carried out a series of zooms, considerable 
differences between the 300 dpi and 2000 dpi resolution 
become obviously evident. 
Galileo Siscam is already developing software which 
allows for the use of the characteristics of the grid of 11 x 
11 reseau crosses, which are present on every photograph 
taken with the Rollei 6006, in order to allow for the 
correction of deformations introduced during a possible 
printing, those of the paper support and those introduced 
by the scanner, as if the total effect were generated by a 
normal distortion. 
2.4 Processing by Orthomap 
The processing by the scanner is called by the 
orthoprojection program with the D.T.M. and the 
corresponding breaklines, after having identified the 
characteristics of the camera and having identified the 
control points with the mouse. 
A large question mark arises in this phase: while 
breaklines in aerial photogrammetry are easily identifiable 
elements, in the architectural field they become a partly 
anomalous element. They can not always be plotted 
because often they are not visible on one of the 
photographs, or on both. For example, a window still must 
be represented by two three dimensional elements, more 
or less coinciding in the view, indicating respectively the 
beginning and the end of the mentioned still. From 
experience one can say that normally only one of the lines 
can be plotted. It is like as if in aerial photogrammetry one 
would have to plot both the eaves and the ground line of 
buildings, without considering that the relative depth 
variations in architectural photogrammetry are usually 
much larger. 
The white areas of the plot are necessarily also present in 
the D.T.M.; there will be parts of the image without 
corresponding 3-D objects, which have known coordinates. 
For these areas, the orthoprojection will be erroneous, or, 
at best, ill determinated, if one does not edit the file of the 
plotted breaklines generating the corresponding lines. 
Apart from this, the program runs sufficiently fast on a 
hardware platforms as that, which was used. 
Figure 2 shows the orthoprojection of the image digitized 
at 300 dpi, processed without using the breaklines, with 
  
  
  
  
  
  
EET 
  
  
  
  
   
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
= 
da m 
| 7 \ 
| \ 
| | x aue o | + © 5 
X | f td ^ 
T | \ +e 
| | D 
| + 
E " 
| E t [ 
md c * 10" * 1 | T 
  
  
  
  
Figure 1 - A longitudinal section of the Santuario. 
    
= AU" ammi Maz uy i VR AAA L4 LU imd mand um fK- —ww-- a AA mnl.