International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
6.2 Automatic Photogrammetric technique
In the last point “5. Mathematical prediction of glacier
dynamic" we have solve the automatism of the "control points".
The position of these points can be predicted through the
previous knowledge of the climatic data and the dynamic of
another campaigns. In doing so, for instance, the even years the
traditional observation techniques can be employed (geodesy,
G.P.S.) and in the odd years, the position of the points can be
estimated with predictive calculations.
In addition, there has been a problem with photogrammetry
because of the convergent geometry. In the near future, the
*Foto-Cartógrafo" software, will be used for recovering the
exterior geometry of the images.
6.3 Previous work to photogrammetric orientations
6.31 Taking photographs: In our case, the photographies
of the glacier are convergent and sloping, and the relation base-
distance is, in many cases of 1/1, that is to say, there is the same
distance between the projection's centers and the object's
distance. Great relations between the base and the distance to
the object makes longer the intersection of the rays and this
allows us to improve accuracy (Figure 4). By contrast, the
conventional analytic and digital stereoplotters cannot perform
the photogrammetric orientations.
| Foto 4 Palo
| E oto
| P16 ‘ +
| a P10
| Foto 5
= 21 a 5 mb Foto 2
pa, =" 7 23
M P11 a8
e? s Opa 24 Wl Foto 1
Foto 6 P1 = 9
pio
& Foto 9
, P4 P7
' Foto 7
Foto 8 P5
P6,
| 100 metros
| Fronte del glaciar
Figure 4. Scheme of the rods, control points or artificial signs of
the photographics takings.
6.3. Premarking control points: In some places like
deserts, beaches, forests and glacier, etc., the natural points
could not be the most indicated targets to be controlled
photogrammetrically. In these cases, the artificial points are
located on land before taking the photographs. Their positions
are specified through topographic measurements, although they
can be foreseen statistically (Sanjosé y Lerma, 2001).
To realize our study in different rocks glaciers (Argualas,
Posets, Veleta) signals of different materials have been made;
for instance, at the very beginning the signals were made of
plastic, but their result was bad because they did not stay
symmetrical from all the attitudes. Later, these signals were
made of cardboard, which have been employed on the
photographs, but their problem was related to traveling. During
last years, the signals were made of cloth, which is very easy to
transport, plus internal rods to give them a squared form.
Figure 5. Design of a signal and its position on a photography.
In order to automate the calculation of the glacier's dynamic,
permanent artificial signals can be left on the glacier (Figure 5).
These signals are made of metal and they are put on the ground.
Following this way, the signal's movement is the same as the
glacier.
7. RECOMENDATIONS FOR FUTURE WORK BASED
UPON EXPERIENCE
Project results have been analysed taking into account the
position and attitude of shorts, quantity of control point and
their distribution:
e Geometry: It is preferred convergent photographs
than normal ones. It should be avoided narrow
parallax angles.
e. Number of photographs: It is important to take care
upon the distribution of photographs, rather than on
the number. Minimum number: three; recommended:
four o five maximum.
e Number of control points: To the orientation, apart
from the control points, another kinds of points can be
observed. In relation with these points:
|l. Employ great quantity of control points does not
imply get better results. The program needs more
than ten similar points, and the system will be
more consistent if all of them are points of object
coordinates.
2. To obtain the same number of photographs, the
'ariation of control points imply errors, although
they are better if there is a greater number of
points.
3. "The medial quadratic error is, in general, smaller
if we only take into account the calculation of
the residual of the referential system of the
control points.
8. CONCLUSIONS
As the geodetic methods only allow us to measure a small
number of control points (targets) on site, and aerial
photogrammetry is too expensive and not ideal for this kind of
projects, close range photogrammetric techniques appears as an
complementary method ideal for the 3D reconstruction and
analyses of rock glaciers (Figure 6). Furthermore, the
combination of multidisciplinary data and methods seems
908
Inte.
prot
glot
The
geor
pred
been
4 cm
This
Actu
and |
helic
phot
after
predi
can
need
grour
consi
Actu
proje
of P
(Sier
volca
Refe
Fabre
las À
Grub
for n
Sensi
Kauf
retree
Natic
Remc
Lerm
Univ
Lôpe:
predi
Cuen
Rame
distri]
Perig
