ibul 2004
not be
rvations
rated by
abilized
. during
aft. The
tortions
re data.
cy data
| nearly
out any
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004
The processing system is able to handle data of all different
HRSC camera types (HRSC-A/AX/AXW) and, provided
adequate interior orientation data is available, of other sensors
based on the multi-line principle. Figure 6 shows an example of
ortho-images and DSM from ADS40 data (raw data kindly
provided by LH SYSTEMS) and of an HRSC-AX data set.
The data of both sensors have been processed with the one and
the same software system, which has been described above.
Note that the parameters and target areas of both flights were
different, thus the data cannot be compared directly.
7. FURTHER EXTENSION OF THE PROCESSING LINE
Figure 7. DSM Improvement of building shapes
(Berlin, 2001, flight altitude 5,000m, subset Adlershof)
top: perspective view using automatically derived HRSC DSM
bottom: perspective view using vector enhanced DSM
Subsequent to the photogrammetric processing line, many
applications based on precise 3D and orthoimage data may
follow. Current developments are addressing also the
411
automatization of frequently required related tasks such as the
extraction of buildings in urban areas and data preparation for
high-sophisticated visualization and animation processes.
Examples are the integration of exterior vector information into
the photogrammetric process in order to optimize the product
charactristics for visualization (Figure 7; Scholten et. al, 2003b)
and the automated enhancement/extraction of linear features,
such as 3D building shapes, in combined ortho-image and DSM
data sets (Mayer, 2004).
CONCLUSIONS
Airborne HRSC system operations and photogrammetric data
processing has proven its feasibility with many applications. It
has demonstrated that the demands on efficient data processing
can be satisfied by means of automated processes combined with
robust and accurate processing based on the multi-stereo
capability owned by multi-line scanners.
Parallel processing on large PC-clusters has been introduced as
a feasible concept that enables handling of very large data sets
and optimal utilization of hardware capacities. It can be a basis
for more commercial applications in the near future.
Nevertheless, there is still a need for additional tools to extend
the product line (e.g. the extraction of linear features and
derivation of 3D object shapes). :
ACKNOWLEDGEMENTS
The contributions of all persons involved in the airborne HRSC
project at DLR Berlin-Adlershof, as well as the co-operation of
Prof. Dr.-Ing. G. Hirzinger and his team at the DLR-Institute of
Robotics and Mechatronics at Oberpfaffenhofen are kindly
acknowledged.