) [|
RADIAL TRIANGULATION, ROELOFS 71 1
| Number Distance | Top. Factor Encre point | Nas Pont
IUE | s | | Triangulation Triangulation [1
Photo- | Flying of | Bridged | Max. Syst. Errors M; q M 14
Scale | Height | Photo- | (kilo- | Aver. Max. 7 yst. Errors, Max. Stand. Errors E
| | graphs | meters) | | | X | y | X | Y
| (meters) | | (meters) | | meter)
| | | |
s 1 40000 4000 i | As 0.10 0,26 057 | 0.51 | 0.24. | 041
" : 60, 6000 15 | 63 014: [036 | 108 027.) 095 | 014
T 1:60,000 5280 20 | 104.5 0.24 | 0.44 | 374 ]os49 |] 080 -]| 082
! 1:60,000 6000 | 20 85.5 0.51 | 1.10 174: |. 270 | 1.27 I; 1.01
aps, 1 : 60,000, 5280 20 | 104.541 0.68 | 100 |: 584 | 824 | 183 | 18
1:60,000| 6000 | 20 85.5 0.88 | 1.70 |; 7.43 |.-951-1 258 |. 194
orm | | | |
reo-
of the propagation of errors in scale- and azimuth-transfer on the final result of the
un- R.T.: the rectangular ground-coordinates of the chainpoints. Anticipating the publication
be of this study in Photogrammetria, some of the results are given in the table. Besides the
ets. data given in the table the following information is needed. The camera inclinations as-
sumed are those of a routine flight; average tilt 15, max. 54'; average tip 23’, max. 40’.
In the case of Nadir Point Triangulation standard errors of 10’ in the determination of
an the camera tilt and tip were assumed. The topography factor in the 5th and 6th columns,
is a quantity to characterize the shape of the terrain: it is the greatest difference in
height in hundreds of meters, occurring within an arbitrarily chosen area of 5 x 5 km.
Thus, a topography factor of, for instance, 0.88 average and 1.70 maximum, represents
a terrain where the greatest height difference within an arbitrary square of 5 X 5 km
tial is 88 meters on an average and 170 meters maximum. Although the table refers to a
single strip — a very fragile construction as compared with a block (see below) — the ac-
ns, curacy is high, which becomes even more evident if the errors listed are converted to
T.; photo-scale.
It is emphasized that the coordinate errors given in the table do not contain the
influence of observational errors. This influence has been investigated by Van der Weele,
who - in a study which will be published shortly — arrives at the important conclusion
in that for wide angle photography the accuracy of scale- and azimuth-transfer with R.T.
ing is higher than with spatial triangulation. It would be interesting to extend this study to
al! give the standard coordinate-errors of the chain-points, due to observational errors. This
would be a nice complement to the data of the table.
as In the mean time the Polish photogrammeter Dmochowski has studied this probiem
be along somewhat different lines [2], which leads to his proposal to apply in Poland, for
tic the greater part a relatively flat country, the numerical R.T. to reduce field work as
m- much as possible. This investigation, which is limited to strips controlled at one end only,
Ley would gain considerably in interest if it would be extended to cover strips controlled in
ia: position at both ends.
As to the computation of the rhomboids, which constitute a chain, there is a very
Ve interesting new approach, made by Ackermann [3]. In lieu of adjusting the directions
measured, he starts right from the beginning with rectangular coordinate computations,
es- using the unadjusted directions. For one of the points of the rhomboid this gives a dis-
er crepancy - closing error — which is distributed simply among the coordinates computed
nt to give the final result. This extremely simple and quick method, which nevertheless is
he in accordance with the theory of least squares, will certainly be a stimulus for the ap-
he plication of R.T.
ice Another computational aspect of R.T. is the connection of adjacent strips, This is
