complex structures and objects. In cases where the 
analysis is not a guiding part of an on-line process or 
too complex to be performed in the same cycle as the 
image acquisition and data extraction, the analysis of 
data is done as a separate phase. In cases where the 
final analysis require the whole set of measured data, 
e.q. generation of a DEM, this is done outside the 
time constrained image generating time cycle. 
  
  
O MNS 
Q MacReflex 
Ll] MapVision 
Bl TrackEye 
1 Im. Acq. 
2 Data Ext. 
3 Analysis 
rj 
O 
Y 
Two cycles 
1 + (2,3) 
One cycle Two cycles 
(1,2,3) (1,2) +3 
  
  
  
  
  
  
fig7 Analysis Cycles 
All analysis in one real-time cycle 
MNS, e.g. Quality Control 
Mapvision, e.g positioning 
Image acquisition and data extraction in real-time 
and analysis in separate step 
MacReflex, manual point identification and 
automated tracking 
Image acquisition in real-time and data extraction 
and analysis in separate step 
Track-Eye, image digitization, manual point 
identification and automated tracking 
  
  
  
  
  
  
  
ot O MNS 
v @ MacReflex 
S St Q [] MapVision 
= 9 @ TrackEye 
o 
ss«I.- 
9 o O 
a En 
o + [1 
© 
| | 
3D e.g. Positioning 
tracking 
  
  
  
fig8 Data Analysis, type and precision 
Comments The precision of the systems are 
approximate and depends of course on the type of 
target point etc. The high speed motion system, #4 
Bl na: a different magnitude of precision depending 
on the different conditions for this application. 
    
  
  
  
  
    
    
    
     
    
    
   
  
   
    
   
   
    
   
   
  
  
    
    
   
    
     
     
  
   
  
  
4.5 Decisions and actions based on analysis 
  
If the task of the system is to guide e.g. a production 
line, the analysis of data should result in a decision 
based on a set of pre-defined rules. The decisions 
made by the system is mainly guiding the actions for 
the actual image at hand, but may also guide the 
future handling of images. 
In the case of a separate analysis phase, the decisions 
made can only affect the processing of the images. 
The image acquisition and formation step can only be 
affected if the whole task is repeated. 
Decisions in one cycle 
Mapvision, e.g. moving an object in the 
assembly line 
MNS, e.g. Quality Control 
5. FOUR IDEAS - FOUR SOLUTIONS - 
FOUR SYSTEMS 
The four systems which are used in the paper to 
illustrate the thoughts are here described in more 
detail. First the ideas behind and questions of special 
interest to this paper and secondly a small technical 
part. The following headlines are used for the 
description: 
a Did you set up any clear goal before starting 
the development, or is the 3D system a 
continuation of earlier systems, e.g. 2D 
measurements ? 
b Do all developing engineers have a similar 
background or do they come from different 
fields? 
c How much do you think the scientific 
background has influenced the system design, 
and if so, which part could have looked 
different? 
d How would you describe the 'photo 
grammetric thinking, if there are any, in the 
data extraction part calibration, orientation 
and 3D calculations 
e How would you describe the system regarding 
Precision 
Reliability, robustness 
Self-diagnosis 
5.1 System 1 MNS, Metronor AS © 
  
a Continuation of earlier 2D system 
b No photogrammetrists from the beginning. 
Background in electrical engineering with special 
competence in CCD arrays.