EN 
  
"oem 
  
  
  
  
Deep DEM TRUE SUR" SEH Urt na 
— 
  
  
The test-field Monti di Revoira 
Strip IH 
— \ | 
t= Crane ence ES 
Le pee Nose le 
| \ an 
e e | e o. V nm 
9 SN. 
mo 
enm. st à Ve : 
  
: . [* ; . : . . ° . ^ a 
| . E \ 
e : | : . ; . X - \ à X-axis » 
: "l . : : + . \ 
Model A A A +B =» . . . . A + B\Model B 
. | . . : ; . > ° *x 
nog oy n "i 
e : . .: s \ ® 
e | : . : . 
+ ; uv 
. . 2 . \ 
  
— — — model limit 
Y" « 
Fig. 3. 
® pass-points Scale j 
0 100 200 300 400 500 600 700 800 900 1000 meter 
 check-points 0 
i 
1000 yords 
81 
8 
g 
8 
8 
2 
3 
8 
8 
equivalent operators and orientation methods, concerning photogrammetrically determined field coordi- 
nates of signalized points and photogrammetric centres representing a good international average. 
6.1 Gross errors. 
The gross errors of the passpoints were detected with the aid of the differences in single passpoints between 
geodetic and photogrammetric coordinates. The evident gross errors (> about 2 meters) were removed at 
a first control. After a new coordinate transformation other gross errors (> about 1 meter) were removed. 
5o those differences were brought down to such small values and with such a distribution that the gross 
errors could not be separated from other errors. (Compare 2.2 and 2.3). 
The obvious gross errors of checkpoints were easily detected and removed. The remaining gross errors 
were removed in the following way. Histograms of the error distribution of every model were drawn, con- 
taining all types of errors. The both tails of these distributions were especially studied. (Compare 5.2). 
They have also been reproduced schematically in the Congress issue of Photogrammetria. In the majority 
of cases the gross errors could easily be separated from other errors. In the remaining cases one could 
hesitate as to the errors in one or in some rare cases in two classes of 10 centimeters width would be as- 
signed to the category gross errors or not. Eventual mistakes in this respect must be very few and further 
have a very little if any influence upon the final result of this experiment work. 
If there was a gross error in some of X, Y or Z, all the coordinates of such a point were removed. No 
distance limited by such a point was used. 
The frequency of gross errors of checkpoints appear from the following table. 
10 
  
70 
76 
22 
20 
20C 
  
re 
Parti 
  
14 
76 
JE 
20) 
[^