3. ASSESSMENT OF METHOD AND RESULTS 
Principally, the filtering methods compare favourably in relation to hill shadings generated by illumination simulation: 
Algorithmically they are simpler and faster. Relief depictions which cannot be explained by means of illumination physics 
like combined hill shadings can easily be generated. Design-based corrections can be formed without the bypass via 
illumination concepts. The cartographer does not have to have a clear conception concerning light- and normal vectors 
etc. The methods permit a smooth integration of elevation contour lines and elevation layers. The relation to the tradi- 
tional cartographic-reprotechnical processing steps is very close. Last not least filtering allows the transition to filter- 
oriented methods of generalisation. 
As far as the disadvantages are concerned, it has to be mentioned that the algorithms only linearly approximate the 
cosine intensities of the illumination functions, which implies that they are numerically more inaccurate. Eckert (1921), 
however, already vainlessly tried to improve the linear hachures of Lehmann (1799) by a sort of "cosine look-up table": 
Fortunately our eyes and brain are relatively tolerant in this respect. Some critical minds might argue that the proposed 
methods are not based on a physical model. Last not least we cannot deny that the power of the filters and the depend- 
ence of an illumination model in some cases lead to surprising but correct effects, in other "serious" but rare cases to 
actually wrong relief representations. 
4. CONCLUDING REMARKS 
The possibilities offered for an optimised relief presentation by image processing and filtering in particular can be viewed 
in an ambivalent way. On one side, the question rises whether all these possibilities are at all conform with the tradi- 
tional rules and concepts of relief representation. It might even be, that a new way of thinking has to take place as far as 
relief depiction is concerned. On the other hand, still, the "classical" tradition of manual relief representation can be hept 
up to a remarkably high extent. 
5. REFERENCES 
Bóhm, R., 1998. Kartographische Reliefdarstellung mittels digitaler Bildfilterverfahren unter besonderer Berück- 
sichtigung kartographischer Reliefschummerungen. Kartographische Bausteine Vol. 14, Dresden: TU 1998. 
Brassel, K., 1973. Modelle und Versuche zur automatischen Schráglichtschattierung. Diss. Zürich 1973. 
Eckert, M., 1921. Die Kartenwissenschaft. 2 Bande. Berlin und Leipzig: de Gruyter 1921 und 1925. 
Hurni, L., 1995. Modellhafte Arbeitsablaufe zur digitalen Erstellung von topographischen und geologischen Karten und 
dreidimensionalen Visualisierungen. Diss. ETH Nr. 11066. Zurich: ETH 1995. 
Lehmann, J. G., 1799. Darstellung einer neuen Theorie der Bezeichnung der schiefen Flächen im Grundriß oder der 
Situationszeichnung der Berge. Leipzig: Johann Benjamin Georg Fleischer 1799. 
Lukas, K., 1994. Evaluation und Weiterentwicklung von Methoden der computergestützten Schráglichtschattierung. 
Diplomarbeit. Zürich 1994. 
Lukas, K., Weibel, R., 1995. Assessment and Improvement of Methods for Analytical Hillshading. In: 17" ICC Confer- 
ence 10" General Assembly of ICA September 1995, Vol. 2, pp. 2231 - 2240. Barcelona: Generalitat de Cataluyna 
1995. 
Messerli and Ives (Ed.), 1997. Mountains of the World: A Global Priority. Parthenon, New York & London, 495 pp. 
Weibel, R., 1997. Digital Terrain Modelling for Environmental Applications - A Review of Techniques and Future Trends. 
In: Geographical Information'97, 3'" Joint European Conference & Exhibition on Geographical Information, pp. 464 - 
474. Vienna 1997. 
632 Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
  
  
1.1 € 
a ba