5. CONCLUSIONS 
In order to improve the delineation accuracy respectively reduce 
the variation width it is of high relevance to have detailed and 
standardised interpretation keys. In addition to that it is 
important to have experienced interpreters to reduce 
insecurities in positioning the floating mark and polygon points. 
Nevertheless a certain variation will persist because some 
border lines are flowing and cannot be fixed to a distinct line as 
well as the individual perceptibility of the interpreters can 
never be totally homogenised. One possibility to take care of 
the persisting variations is the definition of bands of same 
probability instead of lines. But in order to implement this in 
the frame of a GIS a high number of investigations would be 
needed to take in consideration the influence of photo scale and 
complexity. 
6. LITERATURE 
Akca, a., 1984. Untersuchungen über die Anwendung von 
Luftbildern bei der Waldkatastervermessung in 
Entwicklungslándern. Schriften aus der Forstlichen Fakultát 
der Universität Góttingen und der Niedersächsischen 
Forstlichen Versuchsanstalt, Band 78, 1148.. 
Congalton, R, G., 1983. A Quantitative Method to Test for 
Consistency and Correctness in Photointerpretation. 
Photogrammetric Engineering and Remote Sensing 49 (1), pp. 
69-74. 
Glemser, M. 1993. Untersuchungen zu objektbezogenen 
geometrischen Genauigkeit. Salzburger Geographische 
Materialein, Heft 20, S. 97-108. 
Schwidefsky, K. 1964. Die Grenzen von Mensch und Automat 
in der Photogrammetrie. Wissenschaftl. Zeitschrift der TU 
Dresden 13 (1), S. 374-370. 
Zhilavnik, S. 1991. Utilization of Photointerpretation in the 
Determination of Forest Stand Bounderies. Lesnictvi, S. 819- 
829. 
358 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996