Results are simply rectified images and anaglyph images of the
findings allowing a 3D impression of the situation in the field
during the whole process. These images can be viewed with
red-green or red-cyan glasses available for few money. One big
advantage of these techniques is that the persons involved can
see first results already during the field campaign which
improves the trust into the works to be carried out, especially
when new or unknown techniques are used.
For the precise geometric documentation, the whole location is
recorded using standard close range stereo photogrammetry.
The analogue stereo models can be processed using an
analytical plotter. The ergonomics of stereo viewing and the
measurement process, especially for irregular shaped objects
with low contrast, are still better with analytical plotters as
compared to digital photogrammetric stations. In the plan the
outlines of the single bones and height indices of selected points
of the bones give the exact 3-dimensional position of every
single piece. If the outline drawing by itself does not allow the
unambiguous positioning of the finding, the 3D-images can
support this task.
The two locations were divided into several fields which
allowed standardized procedures and reasonable image scales.
The connection between these fields was realized by point
marks on nails designed for use with different image scales and
means of viewing.
FIELD WORK
Equipment. The following equipment was used on site in this
project:
Digital still camera (6 Mega Pixel)
Standard laptop computer
AA inkjet printer
Simple construction level, leveling staff and tape measure
Tripods for level and cameras
Photogrammetric middle format camera Rollei 6006 metric
Software for standard image processing tasks (Corel
PHOTO-PAINT®)
e Software for generating anaglyph images (CRANAG 3D),
anaglyph glasses
e Photogrammetric software (PhotoModeler®)
Preparation. The two locations with findings were divided into
two strips of six fields each and one strip of three fields
respectively, each field with a size of about 1.2 x 1.5 m? and
marked with point markers. These
markers must be usable with analogue
201 and digital images in different image
scales, cf. fig. on the left. The position
of these points is needed for the later
photogrammetric orientation process.
The heights of the markers were determined using the level, the
x-y-position by measuring distances between the points. The
positions were calculated approximately using the simple
formulas for arc section and offset planning in an Excel
spreadsheet. The accuracy of these coordinates is sufficient for
simple image rectification and as initial values for the later
bundle adjustment. The size of these fields was chosen on the
one hand with regard to the image scale for the
photogrammetric measurements, the resolution of the digital
camera, the target scale of the rectified images and on the other
hand, of course, to the situation in the excavation. The position
of the two find locations was measured relative to the covering
roof construction using tape measure and level.
Digital Image Capture. The images were captured
perpendicularly downwards from a simple construction made of
boards from a height of about 3.5 m. As the boards were not
absolutely stable, it was helpful to control the camera from the
laptop especially when the exposure times were long. To enable
the generation of stereo images, every field was photographed
from two positions. A base to height ratio of about 1 : 8 to 1 : 10
was used, which means in this case a base of about 30 to 40 cm.
It is essential that the directions of the exposure axes are
approximately parallel and the distance from the camera to the
object is the same, otherwise the stereo images cannot be
processed with the simple tools as described below. The base
must be selected in the same orientation as the final stereo
images are to be viewed. One image of every stereo pair can
also be used to generate the simply rectified image maps of the
situation. Sufficiently parallel exposure axes can easily be
achieved using simple tools. A tripod should be used in any
case. Simple sliding bars are available as photo equipment
which allow the parallel movement of the camera over a small
distance. These devices are often limited to about 15 cm, which
in many cases is not enough. Another way is to align two of the
three legs of the tripod along a board or another line that is
parallel to the intended base, mark the position of one leg and
then move the tripod along this line for the required distance.
The remaining deviations from the parallel axes are usually
small enough to be corrected sufficiently when generating the
stereo images.
A certain attention should be focused on the illumination,
respectively light conditions in general. For a good stereo
impression, the light conditions for the two images should be as
identical as possible. Thus, the use of flashes connected to the
camera is not recommended, as in this case the position of the
light source is different for the two images which deteriorates
the 3D impression. Using direct illumination or sunlight can
also be problematical, as it often leads to very high differences
in light intensity and very hard shadows. This means that darker
areas in the stereo model cannot be viewed any more. Even if
the resulting images do not look bright or brilliant, it is
recommended to use ambient light in order to get an image
suitable for viewing and interpreting the whole scene. This
might be limited under extreme conditions and the options
available with the camera used. Using high quality lenses is
recommended. The excavation area was covered with a big roof
and the locations below ground level were rather dark. One of
them had to be illuminated using lamps, for the other location
the ambient sunlight was sufficient. The images were taken in a
fixed order along the strips of fields which simplifies and speeds
up the process of image capture.
Analogue image capture. The images for the photogrammetric
measurements were taken with a Rollei 6006 metric middle
format camera on color positive film. The image configuration
for the stereo models of the single fields was similar to the one
of the digital camera. In addition to the images for stereo-
processing taken with a 120 mm lens, images with a 50 mm lens
of the locations have been taken to improve the accuracy of the
calculation of the coordinates of the marked points in a bundle
adjustment.
Taking the photographs took some time. The exposure time was
up to 12 min. for a photo. Reasons are the film material with
low light sensitiveness (ASA 50) and the small lens aperture of
32, which was necessary to achieve a sufficient depth of focus
for the indexed focusing of the Rollei 6006.
—380—
Open Photo:
«Source De
; Photograr
| | Number ©
| Mapping |
c Seale / In
| . (e. mage
(^ Image
; | -
1 Source Ex
| Cot
DATA
Image |
probabl
of the o
not real
result o
not an
type of
marked
referen
scale w
like P
photogt
process
rectific
case the
but dej
accepte
underst
metric
situatic
etc.
In th
Photo
in a sc
images
and the
day. T|
points,
