ul 2004 
MIA 
Camera 
Analyst 
mmetry: 
ler 4.0 
ostation 
romedia 
  
  
esolution 
52x1920 
00x1200 
45x1583 
00x1200 
SERRES 
tric 
(bits) 
tern) 
J 
— á— 
ties 
' Laser 
  
| used 
tion of 
ig and 
camera 
ections 
ges 
in the 
allows 
lequate 
vays in 
quired 
sented 
)scopic 
niques, 
:600 to 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
satisfy the requirements for information extraction at a scale of 
1:100. 
5.4.1.1 Stereoscopic Photogrammetry: planning for 
photographic acquisition is done in a similar way as for the so- 
called normal case (Karara, 1980). Thus, parameters as the 
base, distance to object, and model size are estimated. Their 
estimated values permit to draw the photographic survey plan 
(Figure 4), which also shows the locations of the camera 
positions. 
  
  
  
  
  
Figure 4. Plan of the photograph survey (stereoscopic) 
5.4.1. Monoscopic Photogrammetry: in this case, there are 
a set of empirical rules to draw the plan of the photographic 
survey (Figure 5). For example, the angles between the optical 
axes of neighbouring cameras must be approximately 90°; and 
at least three photographs have to be acquired of the same 
object — one in a central position and two on either side, with 
sufficient overlap to have the same details photographed from 
multiple view points. 
  
  
  
  
  
Figure 5. Plan of the photograph survey (monoscopic) 
5.4.2 Planning and acquisition of control points 
Control points must be distributed in accordance to the 
information extraction technique used. The control points for 
the stercoscopic technique were selected around the limits and 
centre of photographs, whereas those for the monoscopic 
technique were selected in a regular manner around the object. 
The control points may also be artificial or natural. The latter 
should be distinct features of the object, which in our case are 
represented by corners of windows, ornaments and crosses. 
They were measured in a closed polygonal set-up, using a total 
Station with a laser distance measuring device. 
5.5 Architectural Archive 
The complete architectural archive was produced by means of 
the monoscopic convergent technique and a calibrated non- 
metric digital camera. This choice was based on the test results 
presented in section 5.8, which permitted to conclude that it was 
the best choice in terms of quality/price relationship. 
Nonetheless, as the above referred comparative analysis 
addresses the two photogrammetric techniques, the related 
processes are briefly resumed in the following sections. These 
processes are carried out with the ImageStation SSK 
(stereoscopic photogrammetry) and with the PhotoModeler 
(monoscopic convergent photogrammetry). 
5.5.1.1  ImageStation SSK Pro: the stereoscopic images 
were oriented using the standard procedures: interior, relative 
and absolute orientation (single model), or interior and multi- 
photo orientation (block of photographs). As a requisite of the 
software, the images had to be resampled afterwards to the 
epipolar geometry needed to perform the stereo plotting of the 
facades. Stereoscopic photogrammetry is more appropriate for 
highly detailed objects, which is in general not the case for the 
Oeiras mother church. Image acquisition is more demanding 
since the normal case is difficult to maintain in order to cover 
the whole object. Furthermore, a trained operator is required to 
extract the 3D information from the stereoscopic images. 
Therefore, the stereoscopic approach is not as user-friendly as 
the monoscopic one, but it is more appropriate for the detailed 
restitution of the complex shapes so often found on many 
heritage monuments. 
5.5.1.2  PhotoModeler: the approach adopted in this software 
is basically different from the previously mentioned since it is 
based on monoscopic measurements that can be carried out 
using a standard computer. The images are oriented identifying 
common points between 3 to 4 photographs, and related to an 
absolute system using the control points referred in section 
5.4.2. The tools to assess the quality of the solution are used for 
continuous monitoring of the quality of the process while 
inserting new images and control points to plot new details. 
Monoscopic photogrammetry is generally considered to be a 
rather simple technique that can be used by  non- 
photogrammetrists but is appropriate only for the recording 
linear features and details of limited complexity. 
Some results of this processing are shown in Figure 1 (elevation 
of facade), Figure 6 (3D model in vector format) and Figure 7 
(3D model rendered with rectified images). 
amos 
  
Figure 6. Perspective NW of the 3D model (vector) of the 
Oeiras mother church