2004 
  
cy of 
of the 
tional 
metric 
ARSS 
uation 
| data. 
emote 
ystem, 
fice, 
from 
ing of 
ensing 
itional 
ISSN: 
ds for 
sets. 
1-188. 
eri AL, 
unting 
ication 
'hnical 
s J... & 
atellite 
ational 
ods to 
nal of 
2003 
unting 
carbon 
6838 
2003. 
ational 
Cover 
rogram 
nmetry 
81366- 
nov N. 
hen B., 
Myneni 
AODIS 
ing of 
HIGH QUALITY DEMS FOR GLACIER MONITORING - 
IMAGE MATCHING VERSUS LASER SCANNING 
R. Würlánder?, K. Eder^, Th. Geist* 
*" Hermann-Hiller Str. 63 84489 Burghausen — roland.wuerlaender@online.de 
n Photogrammetry and Remote Sensing, Technische Universitaet Muenchen, 80290 Muenchen, Germany 
konrad.eder@bv.tum.de 
* University of Innsbruck, Institute of Geography - thomas.geist@uibk.ac.at 
Commission VII, WG VIL/6 
KEY WORDS: Glaciology, Matching, Laser Scanning, DEM/DTM, Comparison 
ABSTRACT: 
Glacier monitoring is of great importance since glaciers are influencing essential areas of life like water supply, 
energy production, 
flood protection and tourism. Furthermore, the variation of glaciers is a sensitive indicator for local and global climate change. 
One method for the computation of mass balances of glaciers is based on high quality DEMs derived with remote sensing 
techniques. During the establishment of the new Austrian glacier inventory a new a 
pproach for semi-automatic DEM generation 
from aerial images has been developed in order to reach best quality for the DEM with a minimum of effort for manual verification 
and editing. Within the OMEGA project (Operational Monitoring 
System for European Glacial Areas) the Vernagtferner, a glacier 
which is monitored permanently has been covered with airborne laser scanning and aerial images. This way it was possible to 
process both, the aerial images and the laser data and to evaluate both methods with respect to accuracy, economic efficiency and 
overall suitability. The paper summarizes the advantages and disadvantages of both methods for glacier monitoring. 
1. INTRODUCTION 
High quality DEM generation is of great importance for glacier 
monitoring. Aerial Photogrammetry with semi-automatic DEM 
generation by image matching proved to be a suitable tool for 
high quality DEM generation (Würländer, Eder, 1998). The 
Chair for Photogrammetry and Remote Sensing of the 
Technical University Munich together with the Commission of 
Glaciology of the Bavarian Academy of Science has been 
deeply involved in the establishment of a new Austrian glacier 
inventory from aerial images taken in the years 1997 — 2002. 
First experiences have shown, that the generation of the 
required high quality DEM is the most time consuming task. 
Therefore the procedure has been optimised, introducing a new 
approach for analysing the accuracy of the matched points by 
means of a-priori knowledge of the terrain surface. In chapter 2 
the new approach for semi-automatic DEM generation from 
aerial images is introduced. Chapter 3 deals with the 
experiences of laser scanning in a high mountain area, chapter 4 
compares both methods to answer the question "matching or 
laser scanning"? 
2. KNOWLEDGE BASED VERIFICATION OF 
MATCHED DEM 
2.1 Experiences and considerations 
In principle, automatic capturing of DEM by image matching 
should be well suited in glacier regions, as usually no objects 
like houses or trees are found in this inhospitable areas around 
glaciers. On the other hand, the geometric and radiometric 
conditions of high mountain and glacier areas cause a lot of 
problems for DEM generation by image matching methods and 
753 
also for manual point measurements at analytical plotters. 
These conditions are for instance 
— very steep and curved terrain surface, 
— regions not visible in one or more aerial images, 
— extremely rough surfaces like rock or glacier areas 
with a lot of crevasses and 
- low textured areas caused by shadows or snow. 
Experiences with automatic DEM capturing in glacier regions 
(Baltsavias et al., 1996 / Würländer, Eder, 1998) show the 
negative influence of these conditions and lead to the cognition, 
that high quality DEMs for glacier regions can only be earned 
by semi-automatic methods. 
For the new Austrian glacier inventory high quality DEM 
generation has to be done for a total area of about 2.000 km?. A 
large part of this work was carried out at the Chair for 
Photogrammetry and Remote Sensing of the Technical 
University Munich, capturing glaciers in the Ötztal, Stubai and 
Zillertal alps with an total working area of about 750 km? in the 
years 1999 to 2003. Based on the experiences mentioned above 
semi-automatic capturing of DEM from digitised aerial images 
was found as the most efficient and accurate method for DEM 
generation. 
The available matching software is the product TOPOSURF, 
integrated in the digital stereo workstation PHODIS-ST (Dórstl, 
Willkomm, 1994). Verification of the matched points is realized 
by 3D vector overlay within the stereo model and adapted 
measurement routines. 
Carrying out the first part of the work mentioned above, some 
difficulties and lacks for efficient verification have been 
recognized and can be summarised as follows: 
a) The correction of points is much more time 
consuming than a simple point measurement, as correction 
means different actions like 
— identifying the false point,