METRIC OR NON-METRIC CAMERAS 
  
Fıc. 6. Stereocamera in a helicopter taking 
photographs for a forest inventory. (A: Base 
beam, length 4.5 m; B, C: Hasselblad cameras 
500 EL) 
to be most difficult due to the vibrations of 
the base beam of the stereocamera. These 
vibrations were caused by the movements of 
the rotor of the helicopter and air turbulence. 
The camera suspension vibrates like a beam 
supported in its central part. The movements 
are extremely critical if the vibrations of the 
helicopter match the self-frequencies of the 
camera suspension. A great stiffness is neces- 
sary so that the bending of the axis of the 
beam does not exceed the given tolerance, 
which means that the deflection in the cen- 
tral part should remain within 0.3 mm. 
Laboratory tests on a vibrating table have 
shown that a framework in combination with 
shock absorbers is necessary for the camera 
suspension whereas a tube would render less 
favorable results. This effect is due to the 
better damping property of a framework. 
The example shows that it is not useful to 
discuss the metric behavior of non-metric or 
metric cameras alone, but the whole camera 
set-up has to be taken into consideration. In 
  
  
111 
this special case, the application of metric 
cameras would not have contributed to any 
increase of the measuring precision because 
the accuracy limitations originate from a 
completely different source. The require- 
ments for the depth precision were ex- 
tremely low in this case, but this is not the 
rule for photogrammetric measurements and 
very often this is the limiting factor for the 
choice of the photoscale. 
In several investigations the advantage of 
convergent photographs compared with pho- 
tographs taken by stereocameras has been 
pointed out (see Table 2)16.2, The increase of 
precision is considerable and it should be 
noted that erroneous elements of inner orien- 
tation degrade the overall precision less than 
an unfavourable base-to-height ratio. Earlier, 
tolerances for the elements of inner orienta- 
tion were computed. It has been pointed out 
that the test field should include the whole 
object so that the determination of the object 
coordinates can be considered as a sort of 
interpolation. The point determination gets 
more critical if the measurements are ex- 
tended outside the area defined by the con- 
trol points. The deterioration of the point 
accuracy can be estimated with the help of 
the law of error-propagation. The coordi- 
nates of an object point are a function of the 
image coordinates and the orientation ele- 
ments. The variances of these point coordi- 
nates are computed from the linearized de- 
termination equation and the variance and 
covariance matrix of the orientation ele- 
ments. With the covariance matrix taken from 
a camera calibration, these variances have 
been computed for several points within and 
outside the test field (see Figure 8). Accord- 
ing to this estimation the measurements can 
be extended up to a factor of three outside the 
Y 
  
  
  
  
—— — M— — M — — — me NO 
Fic. 7. Factors affecting the measuring precision of distances, in pictures taken by a 
stereocamera. À convergency ofthe two cameras (b) or an erroneous definition ofthe principal 
point (c) will cause systematic scale errors. For a stereocamera mounted in a helicopter it 
proved more difficult to get the camera suspension sufficiently stable and to avoid vibrations 
of the base beam than to reconstruct the principal point in a non-metric camera with sufficient 
accuracy (see Figure 6).