NGE 
ttracted serious 
for the accurate 
nera. In the first 
irgets were then 
2 a network was 
oints were then 
earity condition 
oniwa 1977 and 
nethod and the 
Its obtained for 
ation is given in 
's for the 
| direction 
  
cations 
ghest possible 
est granularity 
1e object being 
otion blur was 
| 2 demonstrate 
d with a faster 
   
   
  
by slow film 
by fast film 
2.3 Signalized targets 
To determine the shape of the signalized target that could be 
more easily identified and pointed in an image, series of targets 
with different shapes were designed (Figure 2). These targets 
were then installed on a surface. Photographs were taken from 
these targets and the potential of identification and pointing the 
target centers were experimentally evaluated. The evaluation 
result indicated that the shapes of the targets are not important 
as regards the pointing precision. However, amongst the targets 
that were recorded with a small viewing angle, circular targets 
indicated better pointing precision. 
To determine the minimum target size, Dm, the following 
relation is adopted: 
S 
Lin =d— (D 
f 
where d is the diameter of the measuring mark; s is the camera 
object distance and / denotes the camera focal length. With the 
view to the analysis described above, circular targets with 
diameters of 2 cm were prepared and installed in appropriate 
places on the body of the Patrol car (Figure 4). To cover the 
entire body of the car, 120 targets were used. The overall 
interval between the adjacent targets was about 40 cm. 
  
r 
Figure 4. Patrol car and the installed targets 
  
2.4 Geodetic network 
The 3D object coordinates of the targets with respect to a local 
geodetic network were determined using the Sokkia Power Set 
1000 total station (Figure 4). The accuracy was estimated to be 
0.1 mm and 1 second for the distance and angle measurements 
respectively. 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
2.5 Photogrammetric network 
The network for the imaging stations were designed based on 
the well known q factor included in the following relation: 
> 28,5 2 
c, (2) 
Jk 
where g. is the overall accuracy in the object space; k is the 
average number of the photographs taken from each station; 
O; denotes the average accuracy of the image point 
measurement and S is the average image scale given by: 
Ya (3) 
where c is the principal distance of the camera. For a 
geometrically rigorous photogrammetric network, q factor is 
considered to vary from 0.6 to 0.7. By taking into account all 
constraints for the photogrammetric network design and with 
several trial and error procedures, the final network presented in 
Figure 5 was designed. 
  
  
  
50 m 
ui (xs 
5.6 
  
  
  
  
  
  
  
  
8.00 m 
Figure 5. The designed photogrammetric network 
2.6 Image coordinates measurement 
To achieve high accuracy for the image coordinates 
measurements, Leica SD2000 an analytical plotter is used 
(Figure 6). The accuracy of the image points measurement were 
evaluated using a reseau plate. The largest error for the check 
points on the reseau plate was about 1.75 micrometer using the 
affine transformation. Figure 7 gives the vector plot of the 
residual errors. 
  
Figure 6. Leica SD2000 analytical plotter