|
|
|
|
|
Figure 3.
THE ANALYTICAL SOLUTION
Figure 4 illustrates the general steps involved in the
analytical solution for the movement vectors. Initially
the comparator coordinates are transformed to a photo
coordinate system. The camera does not have a set of
fiducial marks to define the photo coordinate system.
Moreover, the format of the photo is so large that
neither the edges nor the corners of the photographs
could be measured on the WILD STK-1 Stereocomparator.
Consequently, the x and y-axes of the photographic
coordinate system are determined only approximately
for each photograph.
A precision ruler is used to measure the lengths of
the four edges of a photograph, as well as the x and
y coordinates of two photogrammetric control points.
The x and y-axes of the photographic coordinate sys-
tem are then referenced to these two control points.
The exterior orientation of photo 2 with respect to
photo 1 is then determined using a computer program
that was previously developed for computing the orien-
tation of photographs collected from the LANDSAT
satellite (Wong, 1974). The program is basically a
least-squares solution for three dimensional coordi-
nate transformation.
It makes use of the following basic projective trans-
formation equations:
14
Typical Distortion of
Where xj. y;
Wy, E
0.0
=
= —
— — Y — —
0.0
Glass Plate Under Load
X: C
j Wu. . om. Xy
= C
Y Az | M1 M2 M3 7-76)
Cc
2j May Map Mag Z; - 1
and zj are the photo coordinates of an
image point} Xi, Yi Zi are the object space coordina-
tes of the cor*espünd ilg point in object space; Xj is
a scale factor; X©, YC, and Z€ are the object space
coordinates of the exposure center of the photograph;
and the Wis are functions of the three rotation
parameters'w, ¢ and k. In its present application, the
photo coordinates of the eleven control points measu-
red in photo 1 are used as controls. In addition,
about 35 image points which are located near the out-
side edges of the tunnel model where no particle move-
ment is expected are also used as control points. Thus,
there are about 46 control points used to determine
the orientation of photo 2 with respect to photo 1.
The solution solves for the three rotation parameters
» and K, and three translation parameters X5
of photo 2.
oY
and 2
Z
Using the same computer program, the orientation of
photo 1 with respect to the control points in the
object space was next determined. The solution solves