GRAPHICAL OR NUMERICAL PHOTOGRAMMETRY?, V D WEELE 127 
It should be noted that the foregoing considerations have only a very general charac- 
ter. It has never been proven that it is justified to express the influence of all factors 
occuring in formula 3 in a standarderror. On the contrary, some of them will probably 
not have a random distribution at all. 
Moreover, the fact that the precision of coordinates in a model is not homogeneous, 
has not been taken into account. It is well known that the precision of short distances is 
greater than that of the corresponding point-coordinates, which fact can be taken into 
account by an appropriate correlation factor. 
These details, however, are negleeted because they will have only a minor influence 
on the conclusions to be drawn from the given exposure. 
III. Application of photogrammetry to cadastral purposes. 
IT. 1. 
It depends primarily on the required precision for the cadastral purpose, whether a 
graphical application of photogrammetry is appropriate. 
The answer will certainly be in the affirmative where the cadastre has only a limited 
purpose, such as a purely tax cadastre. In that case it will nearly always be possible to 
choose a photo-scale and a map-scale (coeff. a and b in form. (7) ) which meet the tol- 
erances. 
Considerations which may lead to a numerical procedure, also in this case, could be: 
a. The fact that the cadastre plays a sufficiently important role for the trade in real- 
estate (e.g. by using registration numbers for lots etc.) to impose higher requirements 
for precision. 
b. The same may be true when the cadastral data are frequently used as a basis for re- 
construction of lost or disputed boundaries. 
c. The normal routine to obtain values for areas might be computation from numerical 
data (distances or coordinates). Although these data could be derived from the map 
and prepared for numerical treatment, in some cases the organisation of the admi- 
nistrative work might indicate that it will be preferable to collect these data directly 
at the stage of the restitution of the photographs. 
On the other hand, if m; is great with respect to the other standard errors in the 
formulae (7) and (8) the application of the numerical method will not have much advan- 
tage as far as precision is concerned. In these cases, that is in terrains where boundaries 
generally consist of natural features like ditches, hedges etce., graphical photogrammetry 
may be preferred. 
This preference may be stressed by the easy way in which the number of plotted 
points can be increased so that the lines on the map can be considered more or less as the 
result of an adjustment. The influence of identification errors may thus be compensated 
to a certain extend and the result may be even better than with a simple numerical 
treatment. 
111. 2, 
A numerical procedure may be dictated by the required precision. To have the full 
advantages of the photogrammetric method as far as organisation and economy are 
concerned m; in form (8) should be as small as possible so that the smallest photo-scale 
can be used. This leads to a signalisation of all points to be restituted, before the photo- 
graphs are made. This signalisation should include also all terrestrial points which are 
available or which will be determined during the procedure, and, if possible, a number 
of special points which can be used as control for completion-measurements in these parts 
of the terrain where it can be expected that the photogrammetric restitution will show 
gaps. 
Archives 5