The new computed parameters of interior 
orientation were re-introduced in the multi- 
image orientation program with the known and 
measured coordinates of control points to 
make relative and absolute orientation. Then 
digiting the omologous points in the three 
photograms, it was possible to obtain the 3-D 
coordinates of the real points by a spatial 
intersection 
The comparison between the known 
coordinates and the measured coordinates of 
the signalized control points on the frame 
with three different focal lenght is shown in 
tab. 1. 
We observe that the average value of the 
difference in the three coordinates varies 
between about 8 mm and 32 mm and the 
standard deviation of AX,AY and AZ varies from 
7 to 23 mm. The vector differences are 22 
(for 15 mm), 37 (for 28 mm) and 43 mm ( for 
35 mm), while the corresponding standard 
deviation are 19, 23 and 19 mm. 
Table 1 - Differences between known and measured coordinates of the points of reference 
frame. Values are in mm.. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
à = 15 mm à = 28 MM $235 mm 
POINTS Vector Vector Vector 
AS Ay vaso Balan ay amd lax aye az (3 
1 12 6 2 14 119 2 40. 141 135 [37 [17 52 
2 13 3 2 118 [H4 0 1 {14 6 7 2 9 
3 2 0 4 4 [11 0 20.123 130 [26 5 [37 
i rie 3 2 c Ts 1 Hs 122 1231 {ic (16 129 
5 3 7 1 8 7 2 24 vee [45 [520 2 152 
6 [22 3 2:123 |i? 1 127 180 [27 [ST [28 [46 
7. [12 2 1 H2 [10 1 [42 {45 9 4 2 176 
8 [21 1 2 [21 [10 2 171.152 J96 43 [20 [60 
9. 143. 130 122 [57 S 165 24 {61 [i4 67 
10 [20 271453 7123 *[sa [38 23a. (56 138 63 [13. [5i 
i lee ]27 [ha 172 a fes i34 7? 59 (45 (80 [65 
127] 9 T16 2 116 175 3-135. 138 ]«e [10 [34 155 
13 [iz H3 6 [22 6 2 122 [23 [lii 5 [ie ]e3 
14 [12 4 ;116 ]20. 110 1 à it [37 |a9 1 [54 
NA 1 4 S. 141 0 jai 40 129 |59 [35 
Mean |16 | 8 a lea ji Jia 127 137 1320 lea 49 [43 
Stand. 
Dev 117 ls 6 lie |7 hs fie |es los fig 22 |27 
(*) vector = A A X2 +A Y24A 22 
We can see that the results become worse while for a same image distance the same 
increasing the focal length. 
The result contrasts with expected value in 
relation to the lenses distortion : a reduced 
focal lenght presents usually a major lens 
distortion. In fact the front plane ports 
produce a less relevant chromatic aberration 
and geometric distorsion only in a narrow 
angular field of view on the diagonal ( «40 ? ). 
The maximum field of view on the diagonal for 
the 35 mm is 46?, while for the 28 mm is 59? 
and for the 15 mm is 90°. 
We should hypothize that the effects of 
increasing lens distortion is probably 
adjusted by the fact that the eight known 
points of the frame are contained in a central 
portion of the photo for the 15 mm.( fig.5 ) , 
239 
eight points are located at the border of the 
photograms for the 28 mm and the 35 mm ( 
fig.7,8 ), where the effects of lens distortion 
is more relevant and not negligible. 
The radial distorsion evaluation of the used 
lenses is not the aim of this work, for this 
particular study see Hohle, 1971 and Fryer, 
1986. 
The self-calibration parameter model is not 
useful to describe film deformation and 
unflatness effects because a metric or semi- 
metric camera were not used. But these 
effects in most cases , like this application, 
yield no improvements in accuracy over the 
physical model ( Fryer, 1986).