errors are random in nature, the method of least 
Squares produces the answers of "maximum likelihood" 
according to statistical theory. 
  
  
2.5.1 Measurement Residuals 
  
From earlier discussions it was seen that 
  
a point's location in space may be determined 
by- the. intersection of two non-parallel "rays". 
But,: since more than two photographs are usually 
taken, there may well be more than two rays 
intersecting a point. In the least squares 
  
solution, a single set of xD li coordinates 
are determined for the point and these coordinates 
are, in essence, those that best-fit. all incoming 
rays. Hence, there is only one set of coordinates 
for the point. 
From the preceding paragraph it is seen that 
all rays do not exactly intersect at the computed 
Xp on of the point. For quality control 
purposes it is customary to project a line back 
to each photograph from the computed location of the 
point so as to obtain computed locations of the 
images of the point on all of the photographs on 
which the point was "seen". These computed image 
locations are differenced from the originally 
measured locations (i.e. the x41, y 
jo of Figure 1) 
  
ij 
' 
to provide what are called "measurement residuals". 
With proper work these residuals are typically less