40
Thus, if AB, the relative orientation will be difficult or
0
impossible due to errors in y’, (as well as due to observational
errors in y-parallaxes).
4. The accuracy of the coordinates after the absolute orientation.
In the Appendix it is explained how the residuals dx, dy, dz,
can be derived as functions of the original errors dx, dy, dz.
These dx, dy and dz can now be substituted by their functions
of the basic errors in the inner orientation and in observations
of parallaxes and coordinates. Thus, when the law of propagation
of errors is applied, the weight coefficients of dx, dy and ôz
can be expressed in terms of the weight coefficients of the above-
mentioned basic errors.
dx, dy and dz due to errors in the inner orientation are given
by (9), and introducing (4) into (7), dx, dy and dz due to ob-
servational errors in y-parallaxes are derived.
dx, dy, dz due to observational errors in coordinates, have the
expressions:
=. |
2
dx = — j
dpx + dx
| (10)
dy = — px td)
2
dz= — 7px
where dx is the observational error in x-parallax, and d$ d) are
the errors in identification and setting of the points.
5. Numeral examples and final conclusion.
Tables 1 and 2 are based on following assumptions [1]:
Wide angle photography: c=1,5 dm, b=0.6320 A
d=0,582,
0 5 = Q pps = Q a = Q ist
Mg == 7 HB | (1 1)
m, =m, =m, =20 y
Au ] zo ii
T —9zl1, $20,