International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
  
15 2 >5 [m] 
Figure 6: Distances between the points of the DTM grid and the 
terrain points next to them 
3.2 Interpolation Error dZ in Areas without Terrain Points 
i 
Figure 7 shows a special constellation occurring in areas 
without terrain points: the DTM software bridges big gaps 
between distant terrain points more or less straight. The terrain 
however shows on the left a curvature l/r. This corresponds 
then to an interpolation error dZ in the DTM grid point, with 
the distance d, to the next terrain point: 
next terrain point DTM grid point 
  
terrain point 
Figure 7: Interpolation error dZ in an area without terrain points 
Applying formula (3) with information on local curvature 
(Figure 3), and with the information d, (Figure 6), one arrives at 
results as displayed in Figure 8. The interpolation errors dZ 
vary between -8.52m and 7.12m. In areas of higher point 
densities, with point distances around 1m, interpolation errors 
dZ are negligible. In areas without terrain points, i.c. in areas 
where no statistic accuracy measure could be derived (Figure 
5), the dZ values computed here are of great importance. Where 
the value d, exceeds the radius r of the curvature, no error value 
should be displayed; rather, such areas must be marked as 
unusable (white areas in figure 8). 
Internatio. 
SIC iy 
leads to a 
points. 
The photc 
the ALS-] 
areas whe 
Additiona 
The maxi 
values is 
15cm. Th 
are rough! 
DTM (rea 
to a smoo 
measure c 
2.lem to 
significani 
in Section 
E 
      
02 1 852 imi 
Figure 8: Geometric accuracy of an ALS-DTM, 
white: areas where d, is bigger than the radius of curvature 
4. DTM FROM LASERSCANNER DATA 
For DTMs derived from ALS data, the quality measures as 
described are of special importance. Airborne laserscanning 
delivers very dense sets of terrain point measurements — but 
nonetheless, concerning quality in small areas much is left to 
chance. There are also large areas without terrain points. 
Therefore, in creating ALS-DTMs a subsequent analysis of the 
data and of the DTM derived of them is inevitable. 
Figures l, 2, 3, 5, 6 and 8 are sections of an ALS project in 
Austria. The aim of the project was to model flood risk areas of 
the very flat valley Pulkau. The flying height was about 1000m. : 
An Optech laserscanner was used for data acquisition. 
Figur 
includir 
In the cas 
digitized 
shown in 
breaklines 
2.5¢m anc 
measure a 
70cm to 
(quadratic 
deriving tl 
the breakl 
  
Figure 9: Orthophoto 
S. DTM FROM PHOTOGRAMMETRIC DATA 
In the test area analytic photogrammetric measurements have 
been carried out, too, and a DTM was derived with SCOP. The 
aerial images, using a 15cm camera, have an image scale of 
1:7500. With the practical rule of thumb (Kraus, 2004) this 
116