(2) 
1. Parametric Representation of the Stereoscopic Model 
  
It was shown in an earlier paper (1953) (1) that the coordinates 
of a ground point in the coordinate system attached to the left hand picture 
with origin at its perspective centre are 
( *x,oXy,x) end udo LI 
where x and y are the photographic coordinates divided by the principal 
distance, f, and x is the scale parameter for this point. 
The survey (ground) coordinates X, Y, h are given in terms of 
the coordinates (xx,xy,x) by linear transformations of the form 
RY XX (X) 
YY. =}, xs = HX.R y Eur neni2) 
Z—h | X | Li) 
where Ro is the matrix representing the orientation of the left hand 
picture with reference to the survey coordinate system. 
L mi ni 
R = N m, n, dh oe, (3) 
1; my ng 
This mode of expressing the ground coordinates in terms of the 
photographic coordinates shows that the general affine transformation rep- 
resented by Ro may be applied either to the photographic coordinates 
properly adjusted for variation of scale, or to the raw values. In this latter 
case the transformed coordinates are multiplied by the proper scale. The 
separation of the basic operationsis of some importance, for the scale para- 
meter depends solely on the elements of relative orientation and the paral- 
laxes as measured in the plane of the ‘photograph. It also helps us to 
evaluate the intrinsic precision of the instruments from considerations of 
the way in which they generate the scale parameter, for the affine 
transformation gives no serious difficulty. 
2. Polynomial Representation of the Scale parameter 
  
Any attempt towards the exact generation of x is bound to bring 
about one of the classical types of plotting machines or electronic versions 
thereof. The polynomial representation of this parameter was therefore