QUI- 
  
IDLE 
can 
ns? 
nts 
you 
get 
ro- 
nts 
cal 
or- 
ect 
or 
if 
Our 
get 
m. 
ny, 
Konecny (FRG): 
  
Grün: 
Novak: 
Grün: 
AUTHORS: 
TITLE: 
DISCUSSION: 
Grün: 
  
Proctor (UK): 
  
Your formulation of the equations is very interest- 
ing. Could you not have obtained the same result 
if you had utilized the standard resection formu- 
lation constraining it by equations with appro- 
priate weights? Coordinate differences equal to 
zero,instead of equations for a line, for example? 
Using collinearity conditions. 
The first algorithm which solves the relative 
orientation is based on collinearity. The other 
one uses geometric relations. Of course you could 
try to introduce other conditions, weight them and 
Solve it like any adjustment. But I think it is not 
useful, as you only want to get approximate values 
in an easy way. In the following bundle adjustment 
you can get approximate values in an easy way. In 
the following bundle adjustment you can use weights 
anyway.  - There are different ways to solve this 
problem. I chose the geometrical approach, as it 
is the easiest to understand. 
The question might rise what this paper has to do 
in Commission III, and this question was actually 
put forward. "So 1 want to give you the answer. 
If we accept the area of robotics and industrial 
measurements as a field for our Commission III 
activities as far as models are concerned, those 
papers are of increasing interest in the future. 
A robot for transportation, navigation, etc., like 
a lawn mower or vacuum cleaner - when it travels 
in its environment it sees a lot of edges on build- 
ings and other structures. So, I think as we enter 
these areas this problem of using structured object 
information gets a new meaning. Any more questions? 
Thank you, Mr. Novak, for your paper. 
F. Ackermann and W. Schneider (FRG) 
HIGH PRECISION AERIAL TRIANGULATION WITH POINT 
TRANSFER BY DIGITAL IMAGE CORRELATION 
Thank you very much, Prof. Ackermann. I appreciate 
really your last comment. Are there any questions? 
There should be. 
This is not really a question so much as a comment. 
I am going back at least fifteen years at which time 
there was some experimental work done by the late 
Tony Eder and by Sam Sowton using little spots they 
called star points. They were selected on a mono- 
comparator having very high magnification, and they 
53