Full text: Reprints of papers (Part 4a)

    
    
  
  
  
   
    
   
  
   
   
   
  
   
  
  
  
  
  
   
   
  
   
   
  
  
   
   
   
   
   
   
  
  
   
     
re, the scale of the 
ı and will therefore 
cale (e.g. 1 : 100.000 
‘errors at photoscale 
f 0,02 mm on photo- 
se of this excess of 
> absolute accuracy. 
djustment procedure 
er model allows the 
dinate system, using 
ng the tolerances for 
contain for instance 
ourse depend on the 
E 
ck boundaries 
mental triangulation 
able. The connection 
.. This can be solved 
nly one which takes 
be imposed for a 
described generally 
nly one set of coordi- 
1e camera-axes or by 
f errors are elimina- 
roduced when we use 
je number of models 
ected. 
refore small so that 
ind consequently the 
| the utmost care to 
, etc. In this way the 
    
3 
relative accuracy of the map will conform to previously fixed requirements. The absolute 
accuracy may not reach the maximum which is possible but will anyhow remain within 
reasonable limits. 
Par. 3. Practical execution. 
Fig. 1 gives a scheme of a terrain, covered by a number of parallel strips. The 
boundaries of this terrain in E and S-direction are not indicated since for the following 
description the size of the terrain which could be involved in one adjustment, is not 
important. 
First of all, the strips are triangulated. The results are used for the computation of 
heights (not to be dealt with further) and of X and Y coordinates for each strip. 
  
  
  
  
  
  
  
[EUER 
t Hs n 
C 
SIN v4 7 | 
SN 1^4 dy 
+. Ls 
AS C NN 
—— =) 2 > NN! 
I 
| 
| 
J 
Fig. 3. 
By means of simple linear transformations short parts of these strips are connected. 
The size of the blocks so formed is chosen so that the internal discrepancies remain below 
the value corresponding with 0,1 or 0,2 mm on the map (10 or 20 m on a map 1:100.000). 
For each block two templets are made. Assuming that the common points between 
the blocks are situated in the corners and that these blocks are approximately squares, it 
follows that the most favourable positions for the radial-centres are the mid-points of op- 
posite sides of the squares. This is illustrated in fig. 2 where the circle a is the locus of all 
right-angle intersections of rays from the points A and B. The hatched areas inclosed by the 
two circles b include all points for which the intersecting angle is between 90 +ß and 
90 —ß. It is clear that the greatest part of a square-formed block will be covered by this 
area if the points A and B are corresponding with the mid-points of opposite sides. 
Thus, two templets, forming one block, will appear as illustrated in fig. 3. 
Any connection point will occur in not more than 4 blocks so that it will be covered 
by a maximum of 8 templets. This number of templets can still be handled in practice. 
To allow for freedom of movement of adjacent pairs of templets their respective base 
lines should be perpendicular to each other. A possible base-line pattern is indicated in 
fig. 4. If necessary the portions of the templets indicated by hatched areas in fig. 3 may 
be cut out. 
It is obvious that terrestrial data can also be included in the templets. The distances 
that can be bridged between these terrestrial data may be nearly unlimited since the 
relative accuracy remains unchanged. The loss of absolute accuracy is not important to 
the map-user.
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.