nd is 
used. 
>, and 
1s. Of 
cm. 
photo- 
imilar 
mulas 
ation 
  
  
SYMPOSIUM 
  
NON-TOPOGRAPHIC PHOTOGRAMMETRY 89 
of the quantities b, c, images coordinates and parallaxes are assumed. 
9 
dy eid db D dei drap (6) 
b C bc 
dx = — dy + — dx — — dec (7) 
y x C 
2 2 2 
dz mp da d dsl de 
y Z C 
z z 
= — dy + 4 — ee dc (8) 
y 2 ¢ 
These equations can be changed into different forms, with the aid of (3) — (5). 
For instance, equations (7) and (8) can be converted into: 
2 X 
dim dich dal qo (7a) 
b C bc 
2 , 2 
de = — db + = ds’ — 2 dp 
[^ 
C 
Z Z 
= dll edgy (8a) 
b C bc 
From (6) (7) and (8) it is immediately found: 
1. Because y generally is much larger than x and z, the influence of the errors, 
db, dc, and dp 1s largest upon y. 
2. The most important source of error is dp, since the influence is propor- 
tional to y?. The errors db and dc can be assumed to be very small if the meas- 
urements of the base and the determination of the inner orientation of the 
camera are performed carefully. 
For an accurate niethod for the determination of the inner orientation oí the 
camera, see reference [3]. A regular grid is photographed perpendicular to the 
grid from two stations on the normal from the center of the grid. The correc- 
tons to the preliminary determined data of the inner orientation can con- 
veniently be determined, and also the distortion. Also the standard error of the 
observations can be found in an easy way, and furthermore the standard errors 
of the elements of the inner orientation and the error propagation in functions 
of the elements. 
The next step is to investigate the sources of the error dp. Since p is defined 
as p=x'—x" and consequently d$ — dx' — dx", it is noted that any error of the 
inner or outer orientation that has influence upon the image coordinates x’ 
and x" causes the error dp. Consequently there must be found the connection 
between small errors of the inner and outer orientation on one hand and the 
errors dx’ and dx". Then formula (6) gives the relation between the y-errors in 
control points and the errors of the inner and outer orientation. Knowing this 
connection there can be found the corrections to the elements of the inner and 
Outer orientation from the discrepancies in a sufficient number of suitably 
located control points. 
From the literature, for instance [1], [2], one finds the differential formulas 
for the influence of the errors of the inner and outer orientation upon the image 
coordinates, 
   
  
a 
  
  
  
  
   
  
€———M 
  
  
  
Sls 
    
     
     
     
    
     
   
  
  
EN ik seid ei be