trast to a 
:cially for 
lata man- 
nick data 
DTM and 
rated by 
1the TIN. 
ple" grid- 
me as the 
> "simple" 
yinal con- 
vided and 
reference 
d for the 
h suitable 
  
rid-DTM. 
contours. 
s of finite 
ner/Reib, 
I'M can be 
] skeleton 
  
Fig. 6a: Shaded relief representation derived from a smo- 
othed "simple" grid-DTM. 
lines are used. A hybrid grid/TIN data structure men- 
tioned above is not necessary in this case. To avoid 
problems caused byirregularly distributed contour lines, 
grid points, interpolated from the "simple" grid-DTM, 
can be used as additional reference data (Aumann et al, 
1992). 
To minimize the effort for DTM generation, a direct 
smoothing of the "simple" grid-DTM with the method of 
finite clements was investigated. The smoothed "simple" 
grid-DTM and the smoothed grid-DTM are illustrated 
by shaded relief representations in fig. 6. As one can see, 
both methods lead to similar results. Therefore, the 
smoothed "simple" grid-DTM can be considered as an 
attractive alternative. 
4. CONCLUSION 
The skeleton lines, derived automatically by the method 
shown in chapter 2, supply the additional geomorpho- 
logicalinformation for DTM generation. All three meth- 
ods shown in chapter3 are able to generate a high fidelity 
DTM. The requirement to generate countrywide DTMs 
leads to the conclusion that the TIN-DTM (chapter 3.1) 
is less usefull than the grid-DTM because of the compli- 
cated data structure. The "simple" grid-DTM (chapter 
3.2) is generated by interpolating the DTM grid points 
directly from the TIN. Therefore, the qualityis compare- 
able with the one of the TIN-DTM. Using the method of 
finite elements, either a smoothed grid-DTM or a 
smoothed "simple" grid-DTM (chapter 3.3) can be 
generated. 
Fig. 6b: Shaded relief representation derived from a 
smoothed grid-DTM. 
Finally, fig. 7 shows the contour lines derived from a 
smoothed "simple" grid-DTM. The intermediate con- 
tour lines (thin lines) again demonstrate the high quality 
of the surface description. The example (2,5km*1,5km) 
was calculated automatically and no interactive work 
was needed. 
REFERENCES 
Aumann G., Ebner, H., Tang, L., 1991: Automatic deri- 
vation of skeleton lines from digitized contours. 
ISPRS Journal of Photogrammetry and Remote Sen- 
sing, 46, 259-268. 
Aumann, G., Eder, K., Pfannenstein, A., Würlánder, R., 
1992: Primary data Analysis and Preparation for 
DTM Generation. In preparation for XVII ISPRS 
Congress, Washington. 
Bässmann, H., Besslich, Ph.W., 1989: Konturorientierte 
Verfahren in der digitalen Bildverarbeitung. Sprin- 
ger-Verlag Berlin Heidelberg, New York London 
Paris Tokyo. 
Clarke, A.L., Grün, A., Loon, J.C., 1982: The Applica- 
tion of Contour Data for Generating High Fidelity 
Grid Digital Elevation Models. Proceedings Auto- 
Carto 5, 213-222. 
Ebner, H., Rei}, P., 1978: Height interpolation by the 
Method of Finite Elements. Nachrichten aus dem 
Karten- und Vermessungswesen. Reihe II, Heft 36, 
79-94.