International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 
http://www.ipf.tuwien.ac.at/products/produktinfo/scop/scop_dt 
m. sheet.htm). 
  
  
>0 077 156 282 309 387 464  [polhiynm 
Figure 1: Density N of the original set of terrain points serving 
to derive the ALS-DTM (width of the analyzing grid: 5m) 
2.2 Root Mean Square Error (RMS) 
At the position XY of individual points of the original data set 
("terrain points"), the elevation as measured is compared to the 
corresponding elevation. of the DTM surface. Of these 
discrepancies, per cell of the analyzing grid, RMS (root mean 
square) values will be computed. It represents the accuracy of 
the original elevation values as measured by data acquisition — 
provided that a suitable analyzing grid size is chosen with 
regard to terrain surface characteristics and point density. This 
accuracy 1s varying depending on the quality of data 
acquisition. In case the RMS thus computed in some cell of the 
analyzing grid is smaller than the estimate G4-prioi AS Mentioned, 
it will be replaced by the estimate. Cases of such replacements 
occur in cells with few terrain points in them and/or 1f the 
algorithm applied in interpolating the DTM 1s without filtering 
of random errors (e.g. a triangulation of the given point set). 
Figure 2 visualizes the RMS of the terrain points applied to 
compute a laser DTM with the program package SCOP. The 
maximum RMS value is 52cm; the quadratic mean of the RMS 
values equals 5.1cm. In 64% of the cells the RMS has been 
replaced bÿ O4-prion Estimated as 5cm. The area shown in Figure 
2 and in the figures to follow is identical with the area 
represented in Figure 1. In cells of the analyzing grid without 
original terrain points no RMS is indicated; these cells are 
marked black. 
114 
  
  
0 005 01 02 03 >05 Im] 
Figure 2: RMS of terrain points in relation to an ALS-DTM, 
black: no data available (width of the analyzing grid: Sm) 
2.3 Curvature 1/r 
DTM quality is highly influenced by local terrain curvature. 
Therefore, it is computed for all points of the DTM grid. The 
algorithm applied is based on differential geometry (Appendix 
A at the end of this paper), and it yields the maximum value of 
curvature; figure 3 visualizes these values. They vary between 
-0.61m' and 0.78m ^. The curvature overlay thus computed 
yields a detailed visualization of the geomorphology of the 
terrain (Figure 9 shows the corresponding orthophoto). In 
agriculturally fields, curvature is very low; it is considerably 
larger 1n residential areas, and also in woods. Extreme curvature 
values are typical along break lines. 
    
EEE s 
0.8 0 08 Im 
Figure 3: Curvature values 1/r in an ALS-DTM 
Internatic 
ES 
2.4 Acc 
For estim 
adequate 
procedur: 
individua 
surround. 
plane. T 
computec 
of the nu 
This rule 
where th 
mean; it 
Appendi: 
question 
points. 7 
curvature 
With the 
be compi 
plane w 
specified 
surround 
measure 
points ca 
Fi 
Nume 
that tl 
of in 
howe: 
interp 
proce: