- 12 
© ^ Eres 
This confirms the results of the theoretical studies (Kubik [10] ), that with vary 
systematic image errors present, o 2 will tend to decrease with increasing being 
bridging distance. The r 
block 
b) Instrument: using the same type of camera and with identical photo scales, obtai 
vo 2 (model) varied by a factor 1.2 between the means of blocks measured on a io de 
precision plotter and similar analytical blocks measured by the same organisation | a) Ex 
on à radial triangulator. In an other sample where executing agency, camera type, Qf 
photo scale and bridging distance remained constant, no difference was found ot 
between the average estimators obtained from 8 blocks measured on a stereo 
comparator and 7 blocks measured on a universal instrument. ») as 
ok 
c) Type of camera: in a sample of 14 blocks, where the organisation, photo scale, or 
bridging distance and type of instrument remained constant, a factor of 1.3 2 
was found between the means of o 2 (model) obtained from blocks flown with a 55 
plate camera and those flown with & film camera. Ca 
d) Executing agency: g 2 (model) was found to be remarkably independent of the e 2 e) Ds 
  
executing agency, and no difference was found between the mean estimators 
obtained from samples of respectively 14 and 15 blocks executed by different 
agencies, but with similar camera types, photo scales, instruments and bridging 
distances. The difference between the estimators in fig. 4 must therefore be 
attributed to the influence of other parameters such as type of instrument used, | 
etc. The influence of other parameters could not be established with the | 
sample data available. 
6.3.3. Strips as computational unit (Polynomial adjustment) 
The estimator o 2 determined after a least squares adjustment with a strip 
as computational unit is a measure of the accuracy of the transfer errors. 
In the case of polynomial adjustments, which as can be seen from 4.1, is the 
main type of adjustment carried out, o 2 is a measure of the closeness of fit 
of the correction surface to be control points and additionally in the case of 
blocks, the closeness of fit of adjacent strips. 
A significant difference was found between the estimators o 2(pol.) obtained e o 
after strip and block adjustment, and so the results are tabulated separately. 
The distribution of 02 (pol.) after strip adjustment, obtained from a sample 
of 110 estimators, varied between 2-70 u , with a mean value of 19 u ; 90% 
of the estimators were smaller than 40 u .