7 
photography the relationships take a more com 
plex form in order to account for the sweep 
motion of the camera and the motion of the 
camera vehicle during the exposure. 
In the UAMCE a digital computer calculates 
the photographic coordinates, (x, y), for each of 
the photographs corresponding to a ground hori 
zontal position (X, Y) and an altitude Z e obtained 
by extrapolation from previously measured posi 
tions; the (x, y) values are used to position the 
photographs over the respective scanning units. 
The computer also calculates and outputs the 
partial derivatives of Eq. (2) required to develop 
the appropriate scanning signals. These can be 
written in terms of the derivatives of the func 
tions in the form 
dX 
dx = 
sx 
and 
a dH , v 
dy = 8X dX 
aG 
0Y 
cH 
0Y 
dY 
dY + 
aG 
az 
m 
az 
(§“+§ ■ iï ) 
) 
az 
0Y 
(4) 
Here, dx and dy are displacements in the photo 
graph corresponding to effective scanning dis 
placements dX and dY along the ground; 
~—r = Zx and ~ = Z Y are the components of the 
ax o i 
slope at (X, Y, Z). In practice, dX is a fast sweep 
parallel to the flight line and dY is a slow sweep 
in the perpendicular direction; these combine to 
form a TV-like scan along the ground. The sweep 
components, dx and dy, in the plane of the photo 
graphs produce a one-to-one time correspondence 
between scanning of ground and image points. 
The components of slope are measured by ana 
log circuits that implement the scans described by 
Eq. (4), with independent counters representing 
Z x and Z Y . The signals to change the state of the 
slope counters are derived by comparing the dif 
ferential height error signals on the positive and 
negative X halves of the scanned area for Z x and 
on the positive and negative Y halves for Z Y . 
Because the slope measuring elements form 
closed-loop analog-to-digital measuring units (as 
in the height measurement), the measurements 
are independent of the quality of the imagery in 
the area. The measured digital slope values are 
transferred to the computer for use in estimating 
the altitude for successive elementary compila 
tion areas. 
At each iteration the computer calculates the 
photo coordinates of each new point for the desired 
X and Y and the best estimate of Z available. The 
measurement (by the analog system) of the result 
ing altitude error requires a knowledge of the scale 
factor relating observed displacements in the pho 
tography to altitude errors. In the case of vertical 
frame photography, the scale factor is essentially 
constant for a given model; for all other types of 
photography the scale varies over the compilation 
field. The scale is calculated by the computer as 
often as required to have a sufficiently current 
value for use. The calculations are based on the 
following: 
Let x = G(X, Y, Z), y = H(X,Y,Z) 
x' = G'(X,Y,Z), y' = H'(X, Y, Z) 
be the equations defining the coordinates of a 
point P(X, Y, Z) on each of the photos of the 
stereo pair. Assume that the second (primed) 
diapositive is used to create the orthophoto and 
that its scan is not disturbed by the altitude cor 
rection signals. The problem is illustrated in Fig 
ure 9, which shows an elevation view normal to 
Figure 9. Development of Height Correction Signal, View Normal to Flight Line