The scientific development of photogrammetry has been profoundly influenced 
by the practical application of the discipline being predominantly in one 
field - that of mapping (in its widest sense) from aerial photographs. 
The development of important aspects of the subject can be directly attributed 
to the desire for rapid and economic mapping techniques. An unfortunate, 
but perhaps, inevitable consequence of this dominance is to confuse the 
mapping Application of photogrammetry with the discipline pet 4e. When 
dealing with the different phenomena encountered in many aspects of photo- 
grammetry, this lack of differentiation manifests itself in blurred theor- 
etical concepts together with preconceptions as to the best practical method 
to employ in a given situation. 
Many aspects of close range photogrammetry (certain architectural applications 
for instance) can ‘adopt relatively minor variants of common stereophoto- 
grammetric practice. However, in many fields "standard" techniques may not 
be the most appropriate because aerial photogrammetry is very much a special 
case. In such instances one should reflect on the fundamental characteristics 
Of à general photogrammetric solution, rather than attempting some modifi- 
cation of the topographic "norm". 
The Camera and Stereoscopy 
  
The camera is purely a device which can be used to reconstruct the directions 
from a mathematically defined point (the perspective centre) to a cluster of 
object points. In order to achieve this end the product of the real camera 
is compared with the mathematical standard of the central perspective 
projection. In considering the geometry of close range cameras (especially 
the non-metric type) there can be nothing designed to create more confusion 
than to start from the premise that the air survey camera should be treated 
as the norm. In any case there is a danger "to confuse a standard with the 
mechanism by which that standard was set up" (Thompson, 1957). Photogram- 
metrists should not take the stance of jurors; the ideas associated with an 
infinite fixed focus camera must be guilty of misrepresenting the close range 
case until proven otherwise. For example, what is the physical interpretation 
of the outer perspective centre, and why should it be a unique point (Scott, 
197712 
As the camera can only be used to recreate directions, the determination of 
object point co-ordinates must be achieved by intersection from at least two 
Spatially separated camera stations. This immediately raises the problem 
that, not only must at least two images of a given object point be measured, 
but we must have a means of correlating these images to ensure they relate 
to a unique object point. The widespread use of steroscopic methods in 
aerial photogrammetry has tended to blur the distinction between the 
correlation of an image and the measurement of that image (Thompson, 1970). 
The human anatomy evidently has no place for redundancy in its optical 
Sensors; consequently mapping operations dependent on Stereoscopic correl- 
ation must accept the overlapping pat of photographs as the basic unit of 
measurement. It is indeed fortunate that the requirements of aerial survey 
are readily compatible with those for Stereoscopic measurement, a fact which 
has, perhaps, perpetuated its use into areas more suitably executed using 
alternative approaches. The use of targets, or comparable methods of point 
identification and correlation, allows the photogrammetrist to adopt network 
designs that pay full attention to precision requirements, together with due 
consideration for systematic and gross errors. Of course, in many circum- 
stances point identification or transfer by Stereoscopic means may be 
unavoidable, but even in these cases the use of more than two photographs of 
236