Für die Analyse des Einflusses der geforschten
Erscheinungen wurden durch eine wiederholte Aus-
wertung von mehreren Varianten der Luftaufnahme
die mittleren Lagefehler m,,(s, n) von Punkten des
Versuchsfeldes wie auch der signalisierten Dachek-
ken, und auch die mittleren Lagefehler m,,(o, n)
der nichtsignalisierten Dachecken insgesamt aus 1855
Fallen der Auswertung von Luf ‚hildaufnahme im
Bildmafstabe 1:m, ermittelt.
mx,(s, n) — z 0.015 (cm)
my, (0, n) = = 0.25 | mq (cm)
Auf Grund einer
scheinungen, kónnen
nommen:
Analysis der geforschten Er-
folgende Folgerungen vorge-
|. Die Lagegenauigkeit der photogrammetrisch
ntroduction
When Czechoslovak geodesists investigated why
the planimetric plotting of roof corners was less ac-
curate than that of marked points on the ground,
they found that the corner-to-corner lengths of roofs
calculated from the photogrammetric roof-corner co-
ordinates, were systematically 20 to 30 um shorter
(on the picture scale) than the corner-to-corner
lengths measured directly between the ground pro-
jections of the roof corners.
In view of he great extent of large-scale map-
ping work in the Czechoslovak Socialist Republic,
research into the sources of errors affecting the plani-
metric accuracy of photogrammetry in built
was continued. This research was in particular con-
cerned with the effects of
-up areas
,
the picture scale;
incorrect pointing of the measuring mark;
insufficient identification of the roof tops;
(borders) effects on lines of intense
lighting.
and edge
contrast in
1. Effect of the Image Scale upon Planimetric
Accuracy in Built-Up Areas
(
The effect of the image scale upon the plani
metric accuracy of photogrammetrically plotted roof
tops was investigated on photographs of the settle-
ments Kostelní Stfimélice and (further
"A" and "B", for short). The pictures were taken
in the years of 1962, 19 63 and 1965, with Wild
RC 5 cameras (constants 152/230, 210/180 and 115/
180 respectively) and a Zeiss MRB camera (210
180). The image scales we re such as are usually em-
ployed for 1:1000 and 1:2000 mapping. Settlement
A" is situated at the Creed ne photogrammetric
test field of Pecny.
HAS RE d Le
l'Éenestovice
‘
Approximately 200 roof corners were measured
by intersection at "B" and 117 at "A", for the com-
durch den
auf den
Charakter haben
ausgewerteten Dachecken ist grôsstenteils
Fehler der ide ntifikation der Dacheckbilder
Aufnahmen die einen zufälligen
beeinflusst.
2. Der Einfluss der
ten bei einer Auswertung
nicht festgestellt.
vertikalen Wärmegradien-
von Intravillanen wurde
3. Mit dem wachsenden Kontrast an den Bild-
orenzen der Aufnahmegegenstinde wachst auch die
o = D: =
Lagefehler der nichtsignalisierten Dachecken.
4. Durch die Verbreunungüberreste in den bode-
nahen Luftschichten wird die Sichtbarkeit und da-
mit auch der Kontrast der Bilder vermindert. Man
kann annehmen das dadurch gewissermasse zur eine
Erhóhung der Lagegenauigkeit kommen kann.
r
parative study. The mean planimetric error in the
geodetic measurements of de roof corners was
+ 2.2 cm in settlement À, and
1.3 cm in settlement B.
c
This accuracy is similar to that required for points
of the photogrammetric test field, and the surveyed
oof corners in the localities me ntioned can therefore
field.
Most of the stereopairs were plotted on an SMG
siereometrograph made by Zeiss, Jena, each
being independently plotted three times à certain
time intervals. The geodetical coordinates of both the
test-field points and the roof corners were compared
with the photogrammetrical and numerically plotted
coordinates, and the mea
be considered points of such a
pair
n planimetric errors were
determined. The differences between the mean plani-
I
metric errors derived from pictures taken at “A” and
^B" with cameras of various types and characteristics
were foud insignificant; Table I the
photographic variants are only classified according
to the picture scales
therefore in
The mean planimetric also listed on the
image scale in Table 1, made it clear that the relative
planimetric accuracy increases with the decreasing
scale of the photograph. This fact p rmitted conclu
sions on the effect of. the sy stematic erroi of a con
stant nature,
rrors,
virtually unaffected b nage scale.
derived ne e the values
ál the mean planimetric errors m,, of points numeric-
ally plotted from photographs on a scale of M;
[E [3, 41:
he fotiosins. relations were
Ms A
my,(s, n) ^ = | 5 -] 100 | [em], (1)
: ms
Mxy(O; n) = = 10 + 2.0 1000 —
0.03 ( 30s li
Bi, d 1000 [cm] (2)
