Ali, Mahmoud El Nokrashy 
  
e seen that Table (3) gives the average errors for X, Y, Z and T directions for check points against base line lengths. It is clear that 
imit, then increasing base line length improves the accuracy in object- space coordinates in all directions. The percentage of 
1sed for all improvements in accuracy from using the model of 2.8 m base line instead of 1.8 m base lines are 23%, 24% and 42% 
in X, Y and Z directions respectively. 
3.5 Effect of Object-to Camera Distance: 
To study the effect of the distance between the camera and the observed points on the accuracy of the measurements, a 
the models test on a stereo pair of 2.05 m base line was carried out. The numbers of control and check points were 9 and 12 
| as control respectively. Table (4) gives the errors for each check point against its object to camera distance. It can be concluded 
that the distance between the van and the object point has a great effect on the accuracy. For example the improvements 
percentages in accuracy of a target point at a distance 30 m than that at 53 m are 2596, 2396 and 4096 in X, Y and Z 
  
  
  
  
  
  
  
  
Tage cron respectively. 
maximum 
equal. Table (3) Average Errors Against Base Line Lengths. 
Length of Base Average Error (cm) 
Line ( m) SX SY SZ ST 
1.80 3.73 3.48 7.42 9.00 
2.05 3.54 3.24 6.05 8.45 
2.30 3.39 3.01 6.67 8.06 
2.55 3.20 2.93 6:19 7.56 
2.80 3.02 2.80 5.23 6.66 
3.05 2.89 2.66 4.11 5.69 
  
  
  
  
  
  
  
Table (4) Errors against Object to Camera distances. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
Point Object to Camera Average Error (cm) 
No. Distance (m) SX SY SZ ST 
1 81.9 4.55 4.09 8.92 10.82 
2 78.5 4.34 4.03 8.63 10.47 
3 73.4 4.17 3.92 9.27 10.06 
4 68.4 4.04 3.73 7,89 9.62 
S 62.6 3.89 3.48 7.74 9.34 
6 57.3 3.74 3.31 7.49 9.00 
7 52.7 3.39 3.20 7.24 8.61 
8 45.6 3.23 2.86 6.50 7.80 
9 37.0 3.02 2.13 5.79 7.08 
10 35.4 2.84 2.60 3.55 6.75 
11 31.5 2.67 2.54 5.02 6.23 
12 30.1 2.55 2.39 4.36 5.59 
  
  
  
  
  
  
  
3.6 Effect of Using Additional Parameters 
Since a non-metric camera was used in this study, additional parameters were necessary to be taken into consideration 
to obtain better adjustment results. Additional parameters were used to correct for systematic distortions of the camera 
model. The parameters are two for radial distortion (al,a2) , two for decentering distortion (a3,a4) and two for affine 
deformation (a5,a6). These six parameters were chosen according to Novak (1991) using the following relation. 
; known that 
ordinates. In 
> line length 
tively. Nine 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part B5. Amsterdam 2000. 21