iE 
  
: 
  
MEI Rm 
  
  
   
  
  
  
  
   
  
  
  
lass Es a ML ER Sd à 
neve: 
  
ie Si RR 
  
  
COPS UBER VH 
m 
i 
  
ind 
ET 
d 
US 
v) 
t 
  
  
  
  
  
Fig. 5. Microphoto (scale 7:1) from the marks of the position for a stud in three 
series of inclination. Observe the. small variation within each cornergroup for the 
three series b, g and r. 
vibrate the total assembly. Then all the templets move in the inclination 
direction. All "play" work together in one direction and give pass-point 
positions with rather big errors. Repetitions of the procedure, but with 
the inclination in other directions, give other extreme passpoint-posi- 
tions. The mean of the passpoint-positions for each passpoint is chosen 
as final, and they have an acouracy — according to several tests per- 
formed by the author — which is 3 to 5 times higher than for an ass- 
embly of the ordinary type. 
Empirical investigations 
In order to determine the mechanical accuracy of the MRT and the 
PMRT (precise MRT) several tests have been performed with temp- 
lets of a theoretical shape intended to be error-free. The templet-mate- 
rial was thin (0,12 mm) steel-plates. Due to certain factors in the first 
stages of the trials, the templets got angle errors of the same magnitude 
as should be expécted in templets from photos over slightly broken 
terrain and with normal tip and tilt. 
Summing up some of the experimental results they show that an area 
mechanical radial triangulation with about 1.500 mm between the 
ground controls and 360 X 360 mm templets should have a standard 
error below + 0,5 mm and a maximum error of less than 1 mm. A 
strip triangulation with 10 photos and a base of 150 mm should have 
a maximum error not exceeding 1,2 mm and a standard error below 
-- 0,7 mm. Assuming gyro-stabilized camera or gyro-registration or 
very carefully executed photography the abovementioned figures will 
be lower.