om
of
50,
ted
He
An
da
the
NS.
the
in
ns
nd
the
on,
rk,
nd
he
CIS
he
ity
he
on
nd
Dorffner, Lionel
the possibility to present our results in different media. On the other hand, by establishing a 3-D model and an
information system based on the spatial information of the building, the users of this kind of technologies have the
possibility of working in different scales without losing the relations of the whole structure. The 3D Model of high
quality obtained by photogrammetry will be used as the base data of an Information System, which will contain all
graphic, semantic and raster data belonging to Hagia Sophia. This Information system has the duty to collect all the
information about the building and will be a useful guide for everyone.
4 GEODETICAL AND PHOTOGRAMMETRIC MEASUREMENTS
The complex geometrical form of the dome has been surveyed with the help of photogrammetric methods by employing
the metric camera WILD P31 with color and B/W films. Selecting the right film material in order to obtain the most
natural color became a crucial decision. When working with floodlight that illuminate the dome surface a film for
artificial light is needed. Unfortunately, at fine weather sun light through the top windows of the upper gallery is
predominant causing natural light conditions. The situation again changes as soon as clouds cover the sun. As it is not
possible to work in the evening the optimum choice of films depends very much on the current weather and may be
adapted several times a day.
During a period of 3 days a total of 30 monochrome photographs (using a Kodak TMX 100 film) and 50 color
photographs (30 using an Ektachrome 100S for natural light and 20 using an Ektachrome 64T for artificial light) were
taken of the great dome and the adjoining smaller domes. Finally, for the modeling process 17 color photographs were
chosen. 3 photographs were taken from the floor looking upwards to the dome using a focal length of 100 mm. The
other 14 photographs were taken from the scaffold using the same objective. To cover the whole visible area of about
300 gon the P31 camera was positioned on the scaffold at the center of the dome and after each photograph rotated
about 42 gon. So, 7 photographs were taken at a height of about 43 m. The other 7 photographs were taken in the same
way but at an height of 39 m. This results in 3 stereo models with a horizontal base on the floor and 7 stereo models
with a vertical base on the scaffold.
For image orientation and scaling 68 points located on
the lower gallery as well as on the meridianal arches of
the dome were resected from points of a polygon
established on the gallery of the dome (Figure 2). The
points on the meridianal arches and on the front face
gallery were resected with the electronic theodolite
Pentax ATS 102 from nine polygon stations lying in
the gallery of the dome. All these points are defined in
an overall coordinate system with the origin lying on
the entrance floor of the building. Not to confuse the
similar points on the meridianal arches an other person
signalized them with a laser pointer each time during a
men, Ine gun pe adjusted Figure 2. The measured points on the meridianal arches
SN Iesus E e ec T m: Gi thc om en of the Dome and the Polygon Stations. On the upper
uie Ircridiana arches and the ga tery are Getenmne right part no measurements could be done because of
with a root mean square error of £1 cm. the scattold
5 OBJECT MODELING AND ORTHOPHOTO PRODUCTION
The 10 stereo models were oriented using the 68 control points and evaluated on an analytical plotter Wild BC3. Apart
from the Arabian strokes at the center of the dome and a great number of single surface points also the breaklines at the
meridianal arches were measured in order to get a high quality model of the complex geometrical form of the dome
(Figure 3). In total more than 6000 points were determined by stereo restitution.
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B5. Amsterdam 2000. 175
