Achieved accuracy of the SAS method
The dxa, dya, dza differences between the ground coordinates of
observed points and the photogrammetric coordinates
determinated by the SAS method are presented in Table 1.
Table 1: Differences dxa, dya, dza between ground and
photogrammetric coordinates determined by the SAS method
Detail dxa dya dza
point mm mm mm
1 -5 3 1
4 4 5 3
5 -5 -1 2
6 -6 -2 2
7 0 1 4
8 0 -1 4
9 -3 0 -1
10 -7 3 1
11 1 -2 0
12 0 -2 3
13 2 -2 4
14 1 -3 -3
15 -2 1 0
AVG -2.2 0.2 1.5
RMS : 3.6 2.4 2.6
The achieved accuracy of SAS method was characterized by the
following RMS :
m xa — 3,6 mm m ya — 2,4 mm m za — 2,6 mm (3)
with these systematic errors :
cxa=-22mm cya=0,2 mm cza=1,5 mm (4)
Advantage of these experiment was as follows :
1. The ground survey of control points and the photography
were carried out from the same stations and were synchronized
in one day.
2. Achieved accuracy of the surveying angle measurement
according to (2) and photocoordinates measurements according
to (3) were optimal
3. Stability of the atmosphere was characterised by temperature
within 5 ? to 6 ?C.
3. Determination of the spatial coordinates by the
analytical intersection using the ORIENT system of bundle
adjustment
ORIENT is an universal photogrammetric system of bundle
adjustment which was developed at the Institute for
Photogrammetry and Remote Sensing of the Vienna Technical
University by Dr. Kager
. The ground spatial coordinates and photocoordinates of the
control points were used for the determination of
parameters.The parameters included the following unknowns:
The inner orientation, the coordinates of the projection centers
and the rotations.
Variants of the experiment
In our experiment we determined the inner orientation and
remaining parameters in two variants :
1. Variant OR 1 - ORIENT: The inner orientation was
determined for each photograph, from 5 control points placed
on the border of the dam.This way was used as principal by the
SAS method too.
2. Variant OR 2 - ORIENT: The inner orientation was
determined together for 9 photographs from the stations F3 and
F5 and for 3 photographs from the station F1 (with different z '
h = — 20 mm). 6 - 8 control points were placed in the border of
the photograph.
Determination of the spatial coordinates of control and detail
points was carried out by bundle adjustment for all the
measured points and for 12 photographs.
The height differences between the camera stations are given
in Table 2.
Table 2:Heights and their differences between the camera
stations
Camera Z dz
station m m
Fl 444.5 73.7
F3 370.8 0
F5 425.8 55.0
3.1. Determination of photograph parameters and spatial
coordinates by the ORIENT - variant OR 1
For the parameters determination of variant OR 1 were used 5
control points placed on the border of the dam only. Because
the height differences between the station F1 and F3 is 74 m,
the body of the dam covered only 1/4 of the photograph from
the station F1 (Fig . 2). The similar situation is also for the
station F5.
Photograph parameters
RMS of inner orientation parameters were determined by
bundle adjustment as follows :
mxh-0,002m myh=0,005m mf=0,012m (5)
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B5. Vienna 1996
For the con
values of RI
m x2 —2,01
4. Compar
theSAS me
The basis f
(SAS and (
and Table
(AVG) we
photograph
Systematic
The single
photogramn
photogramn
1).
The photog
photocoordi
orientation
photogramn
errors (AVC
Cxa=-2,2 1
Systematic
Differences
and ground
Table 5) les
cx2=0,1m
The systeme
system com
parameters |
6 - 8 control
Projection
The main p
center in zo
quarter of |
-0 to -15 mi
placed optin
