K. Kubik. The question of signalized pass points is an important one, and requires separate analysis. 
General Discussion 
H. Ebner. I wonder whether the difference in accuracy between the bundle method and others is partly 
attributable to different block size and control density. I think 0°, is influenced by errors in ground control - 
these should be weighted accordingly. 
S.K. Ghosh. There is a saturation of theoretical studies. Accuracy depends on many factors, and I propose 
research into fundamental problems, with the emphasis on practical systems. 
E.H. Ramey. NOS obtain accuracies similar to those quoted by Ackermann. Signalization of control is vital, 
and the configuration of the block is also an important factor. The data analysed in the report was too variable, 
and standardized tests would be preferable. 
K. Torlegard introduced the presented paper “Accuracy of Block Triangulation"' (co-author J. Talts). In 
addition to points noted in the abstract, he found that 0, is dependent on the accuracy of ground control and 
the geometry of the block. 
Friday July 28 1972, 14:00 
Invited Paper: “Theoretical Accuracy Models for Block Triangulation’ by H. Ebner. 
Panel discussion 
F. Ackermann (Chairman). It seems that theoretical studies of the effects of random errors may not be worth 
pursuing further. I suggest that the panel, in commenting on the results, might like to consider further studies 
of computation, and whether there is any evidence to judge the extent to which the conclusions derived from 
theoretical studies are borne out in practice. 
D. Osché. The study is interesting and valuable, but I was surprised that only regular control patterns were 
considered. To what extent will the findings be modified with an irregular distribution as found in practice? 
H. Ebner. With 60% sidelap it should make little difference, also for planimetric control with 20% sidelap. 
However, irregular height control with 20% sidelap will have an effect. 
K. Kubik. I agree that random error assumptions will not yield further information. There is a need to consider 
systematic errors, and study should be concentrated towards the actual errors which occur, and their effects. 
J. Talts. I agree, but I think the sources of systematic error are more within the province of Commission I, as 
they are introduced during primary data acquisition (photography). There should also be concentration on 
finding blunders early, as large blocks are expensive if the adjustment has to be performed several times due 
to undetected blunders. 
H. Ebner. Large blocks are not really a problem. Two presented papers (“The Computation of Large Blocks 
with Anblock and its Application to Geodetic Surveys” by P.R.J. Boniface; and ‘The Planimetric Adjustment 
of Very Large Blocks of Models: its Application to Topographical Mapping in Canada’ by J.R.R. Gauthier) 
deal with blocks in excess of 2000 models, and blunders should be easily found, and will be, if 60% sidelap is 
used. 
General Discussion: 
U. Rauhala. Systematic errors can be reduced by interpolation methods; and the simultaneous adjustment of 
ground survey and photogrammetric observations should give better results, indicate gross errors of targetting, 
and reduce the cost of ground survey by allowing a lower accuracy more homogeneous with 
photogrammetry. 
K. Kubik. The structure of a photogrammetric block can be subjected to a mathematical strength analysis. 
From this point of view, 60% sidelap is a definite safety factor. The extra expense might be justified on these 
grounds. Also the redundancy of observations which this provides can be used to eliminate gross errors. 
J.B. Case. An alternative theoretical approach is to consider the effect of auxiliary data, both as a safety factor 
and for improved accuracy. As for the simultaneous adjustment of ground control and photogrammetry, it has 
some theoretical merit: errors of ground survey co-ordinates are highly correlated in a manner known only if