tion 
ints 
; on 
ute 
wed 
a 
15 
tes system X Y Z supplied by D.E.A. On the otherhand, the re- 
sults obtained in the tests done with the absolute orientations 
on the 12 points of the framework were transformed using a pro 
gram of rototranslation from the co-ordinate system X Y Z to 
the D.E.A. system. In this way, because the calibration had 
been done with the D.E.A. equipment of the car model, the re- 
sults obtained from the various tests (absolute orientations 
on the 12 points of the framework, and absolute orientations 
on the 6 points of the car model) could be compared to the 
list of absolute co-ordinates of the points examined. 
It should be noted that the number of points surveyed with 
the monocomparator was about 40% less than the number surveyed 
on the Stereocartograph, even though the points were chosen par 
tly on the panel and partly on the car body model. 
The results obtained, on the level of mean errors, are 
shown in Table 1. 
A short comment on the transcribed results is necessary: 
as can be seen, the errors in the tests done with absolute o- 
rientation using the 6 known points on the car model are con- 
siderably smaller than those resulting from the orientation do 
ne using the 12 points of the framework: obviously the colli- 
mations done on the pre-marked points have proved to be better 
than those on the pegs of the framework. 
Even though, as was said, we tried to achieve a certain 
degree of homogeneity between the tests done with the various 
cameras, this homogeneity has obviously not been achieved, be 
cause of the different photographic dimensions of the pegs; 
this was noted particularly with the TMK Zeiss, which was put 
at a considerable disadvantage by this; it would be useful to 
repeat the experiments using auto-adhesive marks of different 
dimensions so that the collimation for each camera could be 
done on the most appropriate type of mark and in the best con 
ditions.