tion 
su 
vw. 
the 
bution 
refore 
  
   
a point measured on all three photographs. With these concepts 
in mind, let us consider the results of simulation testing with 
the following relative orientation examples. 
Fictitious data for the simulation test consisted of two photo- 
graphs with essentially the same parameters as in the resection 
problem of the preceding section. The focal length was 150 milli- 
meters, the format 230 x 230 millimeters, and the overlap was 60 
percent. There was a symmetric array of 45 image points in the 
model area. These image points were arranged in a 9 x 5 grid over 
the entire model arza. The standard pattern of six image measure- 
ments was used for the initial relative orientation solution. The 
remaining 39 points were considered to be candidate locations and 
were selected and added to the solution. Figure 4 shows the results 
of this typical relative orientation solution. The triangular 
symbols indicate the 6 points used for the initial solution and the 
Circular symbols are used for candidate points. Five zones resulting 
from the ordered grouping of points by the selection process are 
depicted in the figure. In Figure 5 is plotted the change in the v' 
as additional points are added to the solution. The degree of 
similarity between the results of the resection problem and the 
(1 
relative orientation problem are obvious. This is to be expected 
since the computations for the point selection criterion for the 
resection and model dependent relative orientation are essentially