Grün: 
Helava (USA): 
  
Grün: 
somebody should develope it. Maybe it is on its 
way already in some places. 
a you. First Dr. Helava and then Prof. Torle- 
gârd. 
I made a quick decision here. My company plans to 
announce something in March -87, but I'11 make 
here an advanced announcement, stimulated by your 
presentation. What we have concluded is that a 
good way of doing this is to have an automated 
monocomparator where we can rapidly collect image 
patches on which the points of interest appear and 
then proceed with on-line triangulation parallel 
with point collection. As the patches become used 
they can be thrown away. It is not necessary to 
storage huge amounts of imagery. Even in the case 
of points we plan to use later, we can throw away 
the majority of the pixels. Only a very small number 
of them need to be retained to identify the point. 
A minified version of the image may also be re- 
tained for general location of the point. 
I think we can spend a few more minutes, if it is 
alright with you. We are running out of time now, 
but this is a very interesting topic, practically 
and scientifically, and we should discuss it a 
little more. 
Torlegárd (Sweden): Well, I would just like to make remark on the 
  
Ackermann: 
Torlegárd: 
Ackermann: 
results that were presented here in relation to 
the question of the size of the pixel that was 
raised here. Do I remember correctly that the 
images used were taken on the films Panatomic X 
and Kodac XX. 
About the latter I am not really sure which film 
it was, a standard one anyway. 
It was like that. These are two quite different 
emulsions with quite different values for resol- 
utions and so on. When we remember the results 
they were very similiar when it comes to the trans- 
fer precision. And this might depend on the fact 
that we have the pixel size of 20 microns. And 
the resolutions on the  Panatomic X could be such 
that one could use smaller pixels without loosing 
too much of it and just gaining precision. But 
this is just an idea which crossed my mind. I 
think it is worth while to go on further in the 
direction of finding out where the optimal is. 
Well, I have an opinion about that, because in the 
tests I'll present tomorrow FMC photography really 
got superior correlation accuracy compared with 
standard films, still with 20 microns pixel size 
in all cases. That it did not show up in this 
55