International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
Slopes 
Slopes do show similar problems like highways and dams. They 
can be handled as special areas, but it is better to use break lines 
instead of this. 
Caused by the mirror effect of the water surface, usually there is no 
signal available for water bodies, resulting in missing height 
information about the water surface. But just beside the water, 
height information is available which is sufficient for the height 
definition of the water surface. Typical characteristics of 
watercourses are slopes and also parallel dams. Cuts are not 
causing problems because they are not changed by RASCOR, but 
problems exist with dams beside the watercourse if they have not 
been separated. 
  
result of filtering separately for water bodies 
Figure 6 watercourse 
As it can be seen in figure 9, a global filtering of the whole data set 
is affecting the dam parallel to the water body while the individual 
filtering is not causing problems. The effect of separate handling of 
the water bodies is achieved by the introduction of break lines. 
Bridges 
Bridges do show a sharp change of the height without any slope. 
Of course, if special segments are used just for the bridges, they 
will not be changed because points are only located on top of the 
bridges without more or less variation in height. The break lines can 
include also the neighboured dams, so a handling may be easier. 
Grass ; 
Meadows and willows are simple segments; they do include only 
few points not belonging to the bare surface like located on bushes 
and trees, so they can be handled with the global and the 
segmented filtering without any problems. 
Vegetation 
As vegetation areas a mixture of grass and bushes are defined. 
These segments could be handled in a similar manner like the 
segment grass. Corresponding to this, the segmented as well as the 
global filtering yielded to very similar results. 
Forest 
A global filtering of the whole area was done only with one 
iteration because a second iteration affected especially dams and 
highways. For the forest area a second iteration is recommended. 
The filter parameters are not so sensitive. With a second iteration 
also the global filtering was leading to sufficient results in the forest 
area, but as mentioned, other areas have been influenced. 
  
Figure 7 Results in the forest area 
Empirical results with RASCOR 
study area : Williamstown USA. 
RASCOR can be handled in the batch mode or a sequence of 
batch modes The batch-handling is based on the file rascor.dat 
which may include a switch for the automatic handling as batch job. 
5.2.1. using break lines 
PROGRAM RASCOR 
MAR 2003 
UNIVERSITY OF HANNOVER 
ANALYSIS AND FILTERING OF DIGITAL TERRAIN 
MODELS IN RASTER FORM 
FILE NAMES 
CODE FILE / FUNCTION NAME 
] INPUT FILE grid.dat 
2 INPUT FILE WITH POINT NAMES? Y/N N 
3 OUTPUT FILE OF FILTERED DEM DATA  daxyz.dat 
4 FILE FOR PLOT DATA 
5 FILE FOR EXCLUSION 
8 FILE FOR OUTPUT OF REMOVED POINTS 
9 FILE WITH BREAK LINES williamstown- 
nj.brk 
10 BREAK LINES WITH 1 CODE (1) 2 CODES (2) 
POINT NAMES (P) NO POINT NAMES (N) 2 
FILE NAME = BLANK = NO CREATION OF FILE 
TYPE CODE AND FILE NAME IN ONE LINE 
DEFAULT = NEXT INPUT 
OPTIONS 1 
1 LISTING INPUT VALUES ? Y/N N 
2 LISTING SHORT (S) OR LONG (L) S/L/V S 
3 INTERPOLATION(I) OR ELIMINATION (E) E 
4 OUTPUT WITH POINT NAMES? N/S/R N=NO 
S-NAMES-SEQUENCE R-NAMES-RASTER S 
5 SAME TYPE OF TERRAIN(S) VARYING (V) S 
Internat 
6 FL/ 
OR 
7 NU 
8 US 
9 SM 
FIN. 
11 OL 
12 FÜ 
A=/ 
PRE 
13 SIN 
FILE N 
TYPE 
DEFAU 
Final res 
3845; 
SPECIA 
206. 
END O 
lines : 
Final res 
301 
Program 
the rema 
include 
remainin 
Figur 
Figure €