5 m were interpolated for each epoch in order to support further data manipulation
and analysis. Missing elevation data of areas not covered by the photo flights was
complemented by a DEM (10 Â 10 m grid spacing) provided by BEV.
9.3.2.3 Digital Orthophotos
High-resolution orthophotos with a spatial resolution of 0.5 m were generated using
ImageStation OrthoPro of Intergraph. These were used as a basis for deriving
datasets of lower resolution (1 m, 2 m, 5 m and 10 m) to facilitate image processing
(cp. Sect. 9.3.2.5) and cartographic work. The mosaicing process of stitching
together the overlapping orthophotos was time consuming for the 2003 data since
cloud cover, shadows, strong relief distortion and occlusions forced us to work with
small tiles and to check each tile separately.
9.3.2.4 Glacier Boundaries
Glacier boundaries of the three epochs were interactively mapped as 3D polylines
using the photogrammetric workstation. The glacier boundaries of the Austrian
glacier inventory of 1998 (Lambrecht and Kuhn 2007) were taken as a reference for
the consistent delineation of directly neighboring glaciers. In cases of continuous
snow cover at higher elevations (for 2006), snow accumulation in depressions at the
glacier limits (for 2009) and dense debris cover (for all epochs, Pasterze Glacier
tongue) the delimitation of the glacier boundaries was often only vague or sometimes even impossible. In areas with debris cover we were successful in precisely
mapping the glacier boundaries by superimposing the interpolated contour lines of
the DEM of a younger epoch with the stereo model (cp. Abermann et al. 2010). The
glacier boundaries of 2009 could thus not be checked using this 3D technique,
assuming overall glacier recession.
9.3.2.5 Glacier Flow Velocity
Glacier flow velocity is an important parameter describing the state of a glacier, and
it is also needed for numerical modeling in glaciological research (Oerlemans
2001). Surface flow velocity can be measured by various techniques (Ka ¨a ¨b 2005;
Bollmann et al. 2012). In the present study we applied an image-based technique
based on optical flow estimation.
Kaufmann and Ladsta ¨dter (2003) describe a rigorous photogrammetric technique of how to retrieve a dense field of 3D displacement vectors in multi-temporal
stereomodels using image matching. The authors propose to use pre-rectified image
data, i.e. quasi-orthophotos, for image matching. However, quasi-orthophotos
obtained using accurate and high-resolution DEMs will become ‘true-orthophotos’.
Based on this presumption, the 3D problem can be reduced to a 2D problem. A wide
180
V. Kaufmann et al.
and analysis. Missing elevation data of areas not covered by the photo flights was
complemented by a DEM (10 Â 10 m grid spacing) provided by BEV.
9.3.2.3 Digital Orthophotos
High-resolution orthophotos with a spatial resolution of 0.5 m were generated using
ImageStation OrthoPro of Intergraph. These were used as a basis for deriving
datasets of lower resolution (1 m, 2 m, 5 m and 10 m) to facilitate image processing
(cp. Sect. 9.3.2.5) and cartographic work. The mosaicing process of stitching
together the overlapping orthophotos was time consuming for the 2003 data since
cloud cover, shadows, strong relief distortion and occlusions forced us to work with
small tiles and to check each tile separately.
9.3.2.4 Glacier Boundaries
Glacier boundaries of the three epochs were interactively mapped as 3D polylines
using the photogrammetric workstation. The glacier boundaries of the Austrian
glacier inventory of 1998 (Lambrecht and Kuhn 2007) were taken as a reference for
the consistent delineation of directly neighboring glaciers. In cases of continuous
snow cover at higher elevations (for 2006), snow accumulation in depressions at the
glacier limits (for 2009) and dense debris cover (for all epochs, Pasterze Glacier
tongue) the delimitation of the glacier boundaries was often only vague or sometimes even impossible. In areas with debris cover we were successful in precisely
mapping the glacier boundaries by superimposing the interpolated contour lines of
the DEM of a younger epoch with the stereo model (cp. Abermann et al. 2010). The
glacier boundaries of 2009 could thus not be checked using this 3D technique,
assuming overall glacier recession.
9.3.2.5 Glacier Flow Velocity
Glacier flow velocity is an important parameter describing the state of a glacier, and
it is also needed for numerical modeling in glaciological research (Oerlemans
2001). Surface flow velocity can be measured by various techniques (Ka ¨a ¨b 2005;
Bollmann et al. 2012). In the present study we applied an image-based technique
based on optical flow estimation.
Kaufmann and Ladsta ¨dter (2003) describe a rigorous photogrammetric technique of how to retrieve a dense field of 3D displacement vectors in multi-temporal
stereomodels using image matching. The authors propose to use pre-rectified image
data, i.e. quasi-orthophotos, for image matching. However, quasi-orthophotos
obtained using accurate and high-resolution DEMs will become ‘true-orthophotos’.
Based on this presumption, the 3D problem can be reduced to a 2D problem. A wide
180
V. Kaufmann et al.