341 
Zusammenwirken von Luftbild-interpretation und Luftbeobachtung 
für systematische Waldstudien 
In den schwedischen Staatsforsten werden seit einigen Jahren beinahe alle Forstkarten mit Hilfe 
von Luftbildern und Stereoskop gezeichnet. In Norrland wird diese Arbeit hauptsächlich im 
Zimmer gemacht, nur wenige Bestände müssen im Gelände kontrolliert werden. Es ist doch not 
wendig alle Bestände zu besuchen, weil die Bestandesbeschreibung im Zimmer noch nicht mit 
genügender Sicherheit aufgestellt werden kann. Man hat nun mit gutem Erfolg versucht sowohl 
die Kartierungskontrolle als auch die Bestandesbeschreibung von Schraubenhubern oder kleinen 
Flugzeugen durchzuführen. Nach Versuchen im Jahre 1952 sind bei Forsteinrichtung in den 
Staatsforsten Schraubenhuber während 60 Stunden und Piper Cub während 40 Stunden verwendet 
worden. Dabei sind im Jahre 1953 etwa 7 000 ha, 1954 15 000 ha und 1955 30 000 ha kartiert 
und beschrieben worden. 
Mr. II. Axelson of Sweden, presented his paper on Effect of Photo Scale on 
Use of Aerial Photographs in Swedish Forestry: 
For economical reasons Swedish forestry has to make use of air photos from 
a single photography scale for many different purposes, and for the same reasons 
the scale should be as small as possible. What is then the smallest practicable 
photography scale? 
The Committee of Forest Photogrammetry has made a study of the pro 
blem. Measurements of tree heights were made with a mirror stereoscope and 
parallax bar and various stand factors were determined on pictures magnified 
into 1:15 000 from photography scales 1 :16 000, 1 : 26 000 and 1 : 33 000 (camera 
RC 5a Aviogon 15.3 cm). The study comprised three different areas. 
An examination of the sources of error from measurement with a mirror 
stereoscope shows that the camera tilt may give rise to appreciable errors, even 
in the measurement of tree heights. 
The measurements were made by two trained stereo operators independently 
and resulted in a mean error in the individual measurements varying around 
± 2.5 m (mean height about 15 m). The variations of this mean error are 
surprisingly small, and there is no sign of a constant tendency to greater mean 
errors when measuring on pictures from smaller photo scales. The mean error 
appears to be principally dependent on the degree of clarity of the treetop in 
the picture, the difference in shape of treetops and, of course, on the skill of the 
operator. The systematic error shows a clear underestimation of individual trees, 
whereas the height of the stand is often overestimated. The main cause of the 
systematic error is that an object (a treetop) must have a given minimum di 
mension to be reproducible in the picture, but appears to be likewise dependent 
on the skill of the operator. 
The evaluations were done by three foresters, of whom two had received 
a good training in forestry air photointerpretation. The following factors were 
evaluated: mean height of stands, their density and volume, tree species, age. 
The results of these evaluations do not indicate that the scale of the photo 
graphs had affected the certainty of the evaluation. On the other hand, the 
importance of the forester being experienced in such work, and of his working 
with the greatest accuracy, was strongly apparent. No appreciable difference is 
found in the accuracy of measuring tree heights with a mirror stereoscope and 
parallax bar nor in the evaluation of different stand factors when working to a 
scale of 1 :15 000, whether the photography scale was 1 : 33 000 or 1 :16 000. 
For the purpose for which the aerial photos are usually employed in forestry 
at present, therefore, a photo scale of 1:33 000 appears to be satisfactory. But 
attention must be paid to such practical problems as, for example, the size of 
the picture. For practical reasons small scale pictures should as a rule be magni 
fied, but it is very important, especially for the stereo examination, that the 
pictures are not made too large and difficult to handle. For this reason the 
recommended lower limit of photo scale should at present be around 1: 30.000. 
H. Axelson: Effect of 
Photo Scale on Use of 
Aerial Photographs 
in Swedish Forestry. 
(Publ. VIIA 1)