IV. Experimental Result of a Posteriori Compensation of 
: Systematic Errors and Accuracy Analysis 
ol 1 ) Simulated data 
By using simulated data with random errors of mean sauare 
errors in 5, 10 and 154m and 4 Systematic errors respectivelv 
according to the Single and double models in the block ad just- 
ment with independent models, experiments are made by execution 
of two kinds of interpolated compensation prograns 78 computa- 
tions had been carried out totally, the computational results 
are Shown in Table 1. 
In Table 1, systematic error (1) refers to a theoretical 
fiducial rectanzle distortion, as shown in Fizg.2 
r 
  
  
| 
I 
| 
I 
| 
| 
L 
  
  
s 
B] c—  — — 
BT" wl 
Fig. 2 
Systematic error (2) ‘represents a theoretical fiducial 
lozenge distortion as shown in Fize 3. The above = mentioned 
two computations render mainly models to yield planimetric 
SyStem distortion. 
Systenatic errors (5) in radial distortion, it is chiefly to 
make models to produce heisht System distortion. 
  
  
  
  
7 N Systematic errors (4) show a mixed 
\ \ distortion for radial and theore- 
i \ tical fiducial rectanzle distor- 
M A Fig. 5 tion, they render all three-di- 
à za mentional coordinates in models 
158 
to produce systematic distortions. 
Refined program (1) is the interpolation correction by a 
method of weighted mean values. 
Refined program (2) is the intervolation correction by poli- 
nomials. 
64 = unit weisht mean Square error for planimetrio co- 
ordinates 
- 494 -