H
For?ici- 6.2 C
pator
No jnefer-2 -2 6.20
slo m
2 | S1
le Ss =
3 : | 32
| 0 =
<< see € | ese e
son s 9
| ra |
5 : | ..
| 6.21
i . . J.
| 7 t of th
° - resid
. Te |.
| an es
° 3 d
mi. | 6.22
23 = , i 5 io
: | differ
UN oi | el
SE! | for X
Ü | H uy =
li H Fig. 6. Error distribution of strip III: Camera 2 X RMK 21/18 cm.
TN Picture scale 1 : 6 200—1 : 8 100. The |
1j |
IM 1.3
| E
iu
ah Tabl. 5. The frequency of gross errors of checkpoints.
| 1 | Strip Numb. Gross errors — 0 Gross errors > 0
m and meas. X Y | Z X | Y | Z
iB model points a ; : : : ;
T Numb.| 94 Numb.| % |Numb.| % [Numb.| % |Numb. | 9% Numb. | 9%
|
IA 554 21 3.8 13 2.3 10 1.8 22 4.0 25 4.5 7 1.3
1 B 595 8 1.3 14 2.4 29 4.9 18 3.0 11 1.8 12 2.0
A IL A 1222 53 4.3 37 3.0 19 1.6 40 3.3 54 4.4 24 2.0
i B 963 7 0.7 13 1.3 38 3.9 21 2.2 22 2.3 9 0.9
iM
al III A 940 29 3.1 27 2.9 23 2.4 18 1.9 18 1.9 19 2.0
4 B 900 24 2.7 21 2.3 9 1.0 23 2.6 23 2.6 14 1.6
I6 Sum. 5174 | 142 125 128 142 153 85
E. i Together 775 gross errors of 15 522 coordinates. That means 4.99 9% gross errors.
T || Compare 1 888 erroneous points of 27 411 measured ones at the photogrammetric test Oberriet by the
| H | O.E.E.P.E. That is 6.9 */o errors.
Lu
x The frequency of gross errors thus became about 5 9/o. Anticipating the gross errors to be binomially distri-
|] buted the frequencies of gross errors for samples of different sizes were calculated at the significance le-
"
'* 1l vel 0.1%,
SI
E Al
{ | Number of individuals in a samle | 5 | 10 | 20 | 30 | he
j E | The probability for exceeding the following Z, La
i | number of gross errors in the sample is 0.001| 3 4 5 7
DL. | 6.23
I ] repes
| The cause of the gross errors could be studied only very little. Usual error sources seemed to be mistakes the t
ri at dictation, listening and reading of figures, mistakes at the calculation of means, wrong height scale in
P1] the stereoautograph and wrongly identified points of models. But also other error sources might be there. The
@ | 12
jl
