International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004
nominal plane. The total number of parameters was nearly
doubled, but still it was only a fraction of equivalent num-
ber of a free net model. The new model can represented as
following.
Xi COS 06080;
Vz u Sind,
Z; = 1 - sina; : cos0;
(3)
Also a change was made how the rotation matrix of indi-
vidual camera pose is derived from block parameter an-
gles.
(4)
Rin = Roy, 40.10 ! Ra; 0;
The block adjustment was recomputed with the new model.
Now the systematic pattern was essentially reduced. Simi-
lar illustration of projection center height fluctuation was
depicted as with free net solution (Figure 6). From the
same figure it can clearly be seen that there is approxi-
mately a 6mm height difference between image blocks,
otherwise fluctuation inside one block is rather small in
size of T — 29mm.
1.34 T T T T T T T
Blockl — —
Block IE ------
Ss nm N UN ERR d
uns meet 3
E
=
xS
LL té |
1.325 | 1 1 1 1 1 1
0 50 100 150 200 250 300 350
o— angle (Deg)
Figure 6: The fluctuation of projection center y-
coordinates after adjustment with refined estimation
model. Values are depicted according to two separate im-
age blocks respect to o -angle.
The shift on y-coordinates of the projection centers be-
tween two blocks might have happened when camera was
turned around in opposite direction in order to create the
second block.
4.1 Accuracy assessment
In order to compare photogrammetric data with reference
data the rigid 3D transformation between data sets was
calculated. The length of point differences represents the
absolute coordinate difference between data sets including
inaccuracies of both measuring methods and the coordinate
transformation (Figure 7). The point differences near the
origin are more due to unsuitability of tachymeter measur-
ing in short distances. Whereas the tachymeter coordinates
of points far off are more reliable, the differences between
data sets there are more due to limitation of photogrammet-
ric methods. The second order curve (depicted in Figure 7)
34
was fitted into the data set of point differences. It can be
compared with equivalent representation of simulated data
Internati
in Figure 2. Although, it has to be remembered that the In this
camera model and imaging configuration were not entirely present
equivalent. the idea
techniq
KEG ; final stz
0.09 | - a straig
acl | erized
) to be p
Oy A been te
$ 0.06 | e ur : out any
P wl P J erence |
8 ut to be b
3 1 ] pair me
0.03 + ee y em sen € j is up to
0.02 - SM t | : tr © - is quite
sou © Ted : | Howewvt
2 0 : T 15 20 further |
Distance (m) have to
Figure 7: The length of point differences respect to object ditions
distance. is desig
space f
As mentioned before, the tachymeter data cannot be treated than on
totally free of errors. More representative depiction of how more pi
consistent the photogrammetric data set is, or is not, with terms o
tachymeter data, can be seen in Figure 8. For this represen-
tation, all possible combinations of line segments inside
the data set were calculated and corresponding lengths of
lines in both data sets were compared respect to nominal
values of the line lengths. In Figure 8 the second order Fraser.
curve depicts the prevalent tendency of differences respect topogra
to line lengths. The variation of differences in length is pre- ginia U
sumably due to distance of the point pair from origin. More Heikkir
far off the point pair locates from the origin, more inaccu- Time Ir
rate the point determination is. Based on this assumption ISPRS
it is obvious that longer the line segment is, more prob- A
able it is that at least one of the points will locate farther Heikkir
off. Therefore the variation with longer lines is larger than In: XIX
with shorter ones. In Figure 8 it is essential to notice that Amster
with lines about 10m long the difference is below 10mm
in most cases. This tells about reasonable consistency with Heikkir
data sets. tory. In
Optical
et t SPIE S
0.09 + cd
Heikkin
Put 3 in close
027 - ; ar Long-R
E ae]: & | Greece,
3 0.05 |- . " ubi s i 2 He. d. s A Pe 1 Parian,
p Me T arid | for pan
3 : T der (ed
2 : XXXIV
0.02 F* =
0.01 Ë
un po T un 15 s 20 25
Length of line segment (m)
Figure 8: Comparison of data set point pairwise. The dif-
ference is depicted respect to length of line segments.
