64 LA COMPENSATION DES BLOCS DE BANDES, DISCUSSION 
So that is no trouble at all. 
The block adjustment seems to lend itself 
very well indeed for analytical solutions. I think 
here we have specialists knowing very much 
more about that. 
There are two points I would like to mention 
about spatial triangulation using first-order 
plotters. There is one thing I would like to call 
attention for compensation or adjustability; it is 
a funny word. What I mean is that the smaller 
the total error is, and the more systematic its 
propagation, the better can be the results of any 
adjustment, whatever it may be. If you take 
erratic observations you can apply as many com- 
pensations as you want to and you will not get 
a good result. 
In aerial triangulation observations refer to 
strips. You always have to observe strips, as 
Mr McNair said. Therefore, the most direct way 
of increasing the accuracy of block adjustment 
will be to increase in the first instance the accu- 
racy of your observations. This will give better 
strips. You can do that by using better photo- 
graphs, by using better stereo plotters, reliable 
auxiliary data, statoscope of APR, or anything 
you like, but the main thing is to eliminate 
erratic observations. Jumps and breaks we have 
always had up to now. 
There is another point which concerns 
ground control. It is all very well to think of 
block adjustment by using four points in the 
corners, or anything like that. Those points may 
be 100 miles apart, but you cannot observe 
those points without having intermediate points. 
Sometimes observed points are not good enough 
for block adjustment, as we see it, therefore we 
shall have to use triangulation. Therefore, you 
always have intermediate points at a distance 
of about twenty miles. It seems to me that if we 
can organise the ground control survey in such 
a manner that we can observe groups of points 
rather than one point, where we have the 
primary points, and without those primary 
points you can observe groups without too much 
trouble, particularly if you can use theodolite 
and a tellurometer, or something like that, then 
you can without much trouble get sufficient con- 
trol for, say, three master strips from which you 
can get all the control you need for the filling 
strips. It means you can come back to very 
simple solutions, those which have been men- 
tioned here at a previous meeting, which are 
solutions of adjusting individual strips and after- 
wards adjusting the block. That is what is 
happening still because a lot of organisations 
have to do that. I can tell you that if we have 
well-observed strips using good photographic 
material and statoscope data, the residual errors 
after adjustment are about of the same order and 
nature as the accidental errors of observation. 
That means the strips are adjusted as well as 
they can be. We always have errors of pointing 
and observing, and those you cannot eliminate 
and they form the object of the block adjustment 
itself. 
Mr G. H. ScHUT: I would like to report on 
the results obtained by the National Research 
Council of Canada in its participation in the 
international test on block adjustment. The 
adjustment was an adjustment by means of 
strips, which type of adjustment has been the 
subject of some criticism here. We used the RC7 
photographs, and because of the limited time 
which was available before this Congress met, 
the eight west-east strips only have been trian- 
gulated and horizontal co-ordinates only have 
been adjusted. 
A report on the measurements with a stereo 
comparator, on the analytical strip triangulation, 
and on the adjustment procedure are included in 
the General Report of Commission III. The 
Report also lists the time required for the dif- 
ferent operations, and some figures which give 
an indication of the accuracy of the strip trian- 
gulation and of the results of a preliminary block 
adjustment. 
I do not have to repeat all this, but I would 
like to say a few words about the method of 
block adjustment, and I would like to show the 
final results of this adjustment which were ob- 
tained too late for inclusion in the report of 
Commission III. 
The method consists of the adjustment of 
strips by means of conformal transformations of 
the second degree applied to the horizontal co- 
ordinates of each strip. An iterative procedure 
is used in which the sequence in which the 
strips are adjusted, the choice of control points, 
their weights and the number of iterations, are 
all specified by means of a deck of control cards. 
The iterative procedure and the second degree 
transformation were chosen because a pro- 
gramme for second degree transformation using 
double precision arithmetic — that is, twenty 
decimal digits — had already been coded and 
could be used as a sub-routine in the adjustment 
programme. 
If the errors in strip triangulation cause 
systematic deformations of the second degree 
only, this adjustment gives excellent results. 
However, in practice we will observe that the 
systematic deformations of triangulated strips 
can also contain irregularities which can be 
approximated better by a transformation of the 
thi 
blc 
col 
are 
dec 
we 
Th 
of 
stri 
use 
tes 
gre 
cor 
an 
the 
cor 
trai 
blo 
poi 
pul 
sio! 
Ke 
gr 
  
  
ave 
lap