ı ıternational Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part Bl. Istanbul 2004 
  
A LX 
À X / \ NA, : / VON \ 
E. FAY i WL | 
X { | X fi i 
\ / \ \ ; A \ À 
\// \ f Vh 
\/ Y M \ 
\ \ J \ 
red = W. Kornus DEM purple = P. Reinartz DEM 
Si 3 7 ™ 
í X / NA e f E 
Au NU T Sees IV i 
SEA A / M f \ 4 A Nj Y 
L VN V \ i4 V n / i 
\ A \ A 
\ / NAM 
\ 7 \ \ ^ t 
V \ V S x 
\ \ 
pink = H. Raggam global DEM blue = H. R. local HRS DEM 
^-^ 
^ j 
^ A ft Y 
FX / 
/ M Xs / \ 
\ AJ V \ 
\ i 
green = H.R. local HRG/HRS DEM 
Fig 22: Comparison of DEM profiles with the Reference (in 
black) 
The profiles confirm the DEM visualisation, the W. Kornus 
DEM and the H. Raggam DEM (HRG/HRS) are very close to 
the reference. P. Reinartz DEM follows more or less the 
relief but is a bit smooth and the two last ones (H. Raggam 
global and Hrs DEM) are at the same time very smooth and 
locally very noisy. 
All the results depend on the production processing; 
altimetric biases can be explained by error of modelisation, 
altimetric restitution defects by mismatching processing and 
eventual post-treatment (to fill the matching holes or to 
smooth the automatic results). 
The conclusion of this first part is that the production of 
DEM using HRS data can be very efficient, but the quality 
and the accuracy of the DEM depend on the production 
process and on the kind of DEM which is expected. 
3. TS9: BAVARIA (GERMANY) 
3.1 Input Data 
Location: 12°03'-12°45' E / 47°42'-48°30' N 
Elevation range: 0—1830m 
  
  
  
  
  
HRS data: 
ID scene Year | Month Day 
50622520210011018441S 02 10 01 
50622520210011020162S 02 10 01 
50622530210011018521S 02 10 01 
506225302100110202428 02 10 01 
  
  
  
  
  
Table 23: HRS data on Bavaria 
Reference data: 
Reference data has been provided by the Principal 
Investigator, Dr. Peter Reinartz (DLR, Germany) and is 
composed of the following DEMs (Fig 24) 
In North area : 
DEM "Coarser" N 50, size about 601 x 1001 pixels, 
spacing: 50 m, accuracy: about 2m 
In South area : 
DEM "Laser" S 5 1/4, Nb: 4, size about 1001 x 1001 
pixels, spacing: 5 m, accuracy: better than 0,5 m 
DEM "Laser" S 25 2, size: 348 x 401 pixels, spacing: 
25 m, accuracy: better than 0,5 m 
DEM "map" S 25 1, size about 53 x 401 pixels, 
spacing: 25 m, accuracy: 5 m 
GCP Nb: 81 
  
   
(w ith re contours) 
  
Fig 24: Reference 
3.2 output DTM: 
All the seven following investigators on this test site have 
produced DEM's: 
- PI: Peter Reinartz (DLR, Germany) 
- Colli: Daniela Poli (Institute of Geodesy and 
Photogrammetry, Swiss Federal Institute of Technology, 
ETH Zurich, Switzerland) 
-  Col-2: Karsten Jacobsen (Institute of Photogrammetry 
and Geolnformation, University of Hannover, Germany) 
-  Col-3: Jorge Torres (Division de Fisica Aplicada, 
CICESE, Mexico) 
-  Col-4: Romuald Kaczynski (Institute of Geodesy and 
Cartography, Poland) 
- . Col-5: Alexander Suchkov (Geoinformation Agency — 
Innoter, Russia) 
- Col-6: Konstantin Eremeev (Geo-Nadir, Russia) 
As we had some problems to decipher J. Torres and K. 
Eremeev DEM, we won't give any results from their work 
even if they had also interesting results describe in their paper 
(Eremeev K., 2004, Torres J., 2004). 
The remaining five investigator DEMs have been analyzed as 
mentioned in table 25. 
    
Inter 
  
In s 
kind 
mod 
(Pol 
and 
2004 
AS 
sam] 
fine: 
samt 
the : 
area: 
also 
area 
3.3¢ 
All 1 
for ¢ 
refer 
diffe 
341 
3.4.1 
spac 
360 
  
  
  
We 
rece 
20m 
so it