International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part BS.
1. INTRODUCTION
This paper sets out a new method using photogrammetry in
order to survey a scene engraved on a non-plane surface.
This study comes within the context of research conducted
jointly by the GRCAO (Computer Aided Design Research
Group) of the University of Montreal and by the laboratory
MAP-PAGE (Photogrammetry and Geomatics Group) of the
INSA Strasbourg. It has been carried out within a larger project
called « Computer modelling as a means of reflection in
archaeology : a new approach to epigraphic and architectural
survey applied to the Karnak Temple». The GRCAO is in
charge of this project through the Karnak Hypostyle Hall
Project (KHHP) of the University of Memphis, and the
collaboration contract signed by the two institutions put all data
and material linked to the project at the GRCAO's disposal.
One of the project's issues is to survey all the inscriptions
engraved not only on the walls, but also on the 134 gigantic
columns of the Hall. As an aim is notably to publish the
hieroglyphic texts, the method to be developed should have a
two-dimensional result.
Therefore, this study proposes a means to survey hieroglyphic
engravings of columns in a two dimensional environment.
Computer tools have been developed that allow the epigraphists
to draw scenes and record the drawings with their meaning for
future exploitations. Interestingly, it is based only on a series of
snapshots of the column's surface. Thanks to the partnership
above-mentioned, these tools have been eventually tested on a
column of the Karnak Hypostyle Hall.
The survey method can be broken down into two main phases :
the three-dimensional reconstitution of a column, and then the
development of its surface for the two-dimensional surveying of
the epigraphy.
The first part of this paper deals with the construction of a
vectorial 3D model of a column. Three aspects have to be
considered : the data acquisition strategy, the data available
beginning the study, and the layout of a theoretical cone that
represents the column.
The second part concerns particularly the epigraphic surveying
of the column. There are also three steps : the development of
the previously obtained cone, the orthorectification of the
column's original photographs and finally the montage of these
orthophotographs in order to draw the hieroglyphic signs in.
2. THREE DIMENSIONAL RECONSTITUTION
OF A COLUMN
Through photogrammetric formulae, a vectorial three-
dimensional modelling of a column has been calculated from
photographs that cover the entire surface of the column, and
from control points scattered proportionally on its surface.
2. Data acquisition strategy
The problem of the survey in the Karnak Hypostyle Hall stems
from the fact that there are 134 columns erected in an area of
only 102 x 52 square metres. À consequence is an obvious lack
of space, since the diameter of the columns measures 2.8 m,
while the space between them varies between 2.6 and 3 m.
Moreover, the height of the columns is also a problem to reckon
with, since 12 of them measure 21.20 m (in the centre of the
hall) and the other 122 measure 13.17 m. (Figure 1)
Figure 1. Columns of the Karnak Great Hypostyle Hall
These constraints must be taken into account when choosing a
strategy to take snapshots of a column in situ. Eight
photographs of the column could be taken (for one level), in
order to insure a good coverage of its surface and a sufficient
overlapping of the photographs. The positioning of the different
photographic stations within the Hypostyle Hall could be as
shown below. (Figure 2)
EJ Do a column
Figure 2. Position of the photographic stations
Considering that the greatest distance in space of the camera
from a column is about 7 m, and that each of these structures is
very high, several snapshots are necessary to cover its entire
vertical surface. In order to cover a surface ranging from 0 to 6
m, a picture must be taken 3 m away from the column, while 9
m are necessary to cover between 6 and 12 m. Thus, 16
photographic stations are at least needed to cover the whole
surface of the smallest columns and 24 for the biggest ones.
Such a number of photographs may appear constraining, but
one must be aware that the epigraphist takes them with only one
camera using the same lens. Moreover, the photographic
coverage is practically the only work to be done in situ, because
the coordinates of the control points can be calculated later
through Bundle Adjustment. No important topographic survey
is thus necessary, since the created independent model will be
scaled down to proportion by measuring a length on the column,
or will be replaced in the 3D space while surveying just three
points in the field.
2.2 Data available at the beginning of the study
The only data required to create a 3D model of one part of the
column are the eight photographs covering its entire surface and
some control information to be located on these photographs. À
minimum of six homologous points between two consecutive
overlapping photographs are required, and three more common
points between three overlapping photographs. This has been
done using a graphical software, in this case AutoCADO
because it is widely in use in archaeology. The photographs
with the points can be seen below. Polylines that connect the
different homologous points between the shots have also been
drawn for a better visualization. (Figure 3)
Istanbul 2004
In
wl
Th
Fo
