M E | 
( facts ) | 
S e uu m | 
| 
| 
VETT SEEN. E I NFERENCE 
( relationship ENGINE 
“eg rules ^ 
KNOWLEDGE 
BASE 
  
  
  
  
(USER INTERFACE 
  
/ data lor 
( current 
N 
PEE 
/ qe 
\ 
hypotheses 
/ 
  
  
   
WORK! NG 
MEMORY 
  
  
  
Fig. *1 Structure of an expert system 
USER DATABASE FILES 
ona ECT | | Cameras.dbf | 
Table 1: Result of consultation session for three cameras (o, in mm) 
Geometry Zenzanon P31 P31 * UMK 
[55x55] [100x130] [100x130] [130x180] 
n $ £ = 150 mm | £ = 100 mm | £ = 100 mm | f = 100 mm 
o, = 2.4 og; = 3 um 9, = 2 um 90, = 2 um 
um 
2 45 0.310 0.213 0.142 0.109 
3 30 0.254 0.174 0.116 0.089 
4 22 0.220 0.151 0.101 
5 18 0.196 0.135 0.090 
6 15 0.179 0.123 
7 13 0.166 0.114 
8 11 0.155 0.107 
9 10 0.146 0.101 
10 9 0.139 0.095 
11 8 0.132 0.091 
12 8 0.127 
* The database entry uses 9, = 3 um. We include 9, = 2 um for comparison. 
         
     
  
  
  
1 
station 2 
à (X92. Y92.Z 93] 
object-point i 2 3 
io Vs es p 
i i 
p 
4 
5 
Fig. 2: Symmetric network geometry 
[optconti .abe 
  
     
    
    
   
    
   
  
* dimension 
* required Or 
  
  
  
E cL 
| KNOWLEDGE-BASE | 
Fig. 3: Main components of 'camera selection' module 
* name 
* focal length 
  
  
  
  
n 
$ 
e 
**x 
rror factor 
  
  
  
     
  
* rules 
  
* accuracy predictors 
  
  
Y 
  
  
Decisions: 
* Camera 
(suitable or unsuitable) 
* Initial network geometry