8 ANALYTICAL AERIAL TRIANGULATION, AUTHOR'S PRESENTATION 
  
  
now been abandoned by the Ordnance Survey. On the other hand, although the succes- altho 
sive modification of the coefficients has been tried by the Ordnance Survey (Arthur id 
1959), the opinion there is that it does not result in any saving of effort. J e: | 
References. phot 
Arthur, D. W. G, 1959 Photogrammetric Record, III, 14, pp. 112-124. in m 
Thompson, E. H., 1944 A Method of Determining Relative Tilt, Air Survey comp 
Research Paper No. 4, The War Office, London. deter 
1950 Empire Survey Review, X, TT, pp. 330-332. or Nc 
1956 Photogrammetric Record, 11, 8, pp. 145-150. and c 
1959a Photogrammetrie Record, III, 13, pp. 55-59. T] 
1959b Photogrammetric Record, III, 14, pp. 152-159. tion ‘ 
1959e Photogrammetria, XV, 4, pp. 163-179. sayin 
condi 
SUMMARY. : 
relati 
This paper raises certain problems in the computation of analytical aerial trian- rays 
gulations in such a form as to stimulate discussion. The subjects raised are: the treat- such 
ment of redundant observations; the incorporation of models into strips and blocks; the . 
condition equations for relative orientation; the use of an intermediate stereographic COD! 
projection; and the iteration process. pect, 
I thir 
RESUME. tion. 
Le présent article met en avant certains problémes sur le caleul d'aérotriangulations In 
analytiques de facon à stimuler la discussion. Les sujets mis en avant sont: le traitement of wI 
des observations superflues; l’incorporation de modèles en bandes et en blocs; les équa- clusic 
tions de condition pour orientation relative; l’emploi d’une projection stéréographique meth 
intermédiaire; et le procédé de réitération. of ke 
all th 
be th 
Author’s Presentation of the Paper relati 
have 
In the short time I have this afternoon, I do method of dealing with the adjustment of blocks. of ar 
not wish to do more than to draw your attention Here one can go the whole hog and one can there 
to five points which I have made in my paper treat all the observations together and have no the n 
and which I hope might give rise to some dis- approximations; one can deal with the thing in a grapl 
cussion. massive way. However, this is not necessarily Co-or 
In the first point I make in my paper I draw the way that would be the most economical and for e: 
attention to a fact which seems to have been in geodesy, as I point out in my paper, there unkn 
overlooked and that is that in the calculation of are various ways in which this can be and is sim- Mess 
analytical aerial triangulation there really are plified. For example, it is unusual in geodesy whicl 
two problems. There is the problem of deter- not to adjust the station observations first and on d 
mining the unknowns to an accuracy within accept them before doing a figure adjustment. the ; 
shouting distance, shall we say, of the areas of Again, it is not an uncommon procedure, for Very 
observation and then there is the second prob- example, to adjust internally a figure without the e 
lem of introducing redundant observations and introducing a fixed point or conditions which not t 
improving the solution in whatever way is con- are then introduced later on. All these devices simp 
venient or possible. while being theoretically incorrect are never- T 
These two problems are essentially different, theless very important in that they result in a the p 
because one is a non-linear problem, if one has considerable amount of economy. The same proje 
no idea what the unknowns are then the prob- 
lem is a question of solving non-linear equa- 
tions, whereas the second problem is a linear 
problem, although as I said it may be very com- 
plicated, nevertheless it is a problem of the 
solution of linear equations; and these two 
things should be distinct. I do not think they 
are kept distinct either in the literature or in 
people's minds. 
The second point I raise in my paper is the 
thing can be done in photogrammetry as, for 
example, by Jerie’s method, in which he regards 
a section or two models making up a section as 
being a rigid figure which is not to be altered. 
This is exactly equivalent to the method that is 
used in geodesy for adjusting, by variation of 
co-ordinates, a triangulation in which the angles 
measured at a station are accepted and the 
angles between the rays accepted and not 
further considered. So that is not a new idea,