5.4 Georgian Bay (ADS80)
Located on the coast of Lake Huron’s Georgian Bay in Ontario,
Canada, this block is dominated by forest. The imagery has
been captured by North West in 2009 for the Ontario Ministry
of Natural Resources (OMNR) in a GSD of 30 cm.
The resulting info clouds as shown in Figure 8 have already
proven their value to OMNR for extracting forest inventory,
based on the ADS’s distinct, calibrated color bands in combina-
tion with the high point density that exceeds the resolution of
LiDAR data available in that area. Individual tree crowns are
visible in the info cloud; the additional spectral information
allows for manual extraction and/or automated classification of
different tree species.
5.5 SGM Performance
The performance of SGM depends predominantly on the dispa-
rity range, which is determined by the terrain on one hand and
sensor and imaging configuration (GSD and base to height
ratio) on the other hand. This correlation can be clearly seen in
the performance numbers of Table 2.
GSD | Max. | Perfor-
Sensor Data Set Dispa- | mance
[m] rity [pts/s]
Georgian Bay 0.30 104 58,700
ADS80
Sacramento 0.15 488 34,200
DMC-II 140 | Aalen 0.05 296 47,800
RCD30 Bregenz 0.15 96 71,400
Table 2: SGM processing job statistics for different sensors and
data sets.
The vast majority of time in our CPU implementation is spent in
the SGM cost computation and aggregation, which is identical
for ADS and frame imagery. Considering that also the pre-pro-
cessing and the final ground projection are similar operations,
almost identical processing times can be expected in the same
terrain and with comparable imaging constellations, i.e. frame
overlap with a base to height ratio similar to the (fixed) ADS
stereo angles.
6. CONCLUSION AND OUTLOOK
The paper presents our SGM-based DSM extraction software
that can be used with different types of imagery, including line-
scanner and frame data. The ADS implementation has been
used extensively in production at North West since more than
two years, and it has become part of the Leica XPro ADS
ground processing software. The extensions and adaptations to
frame imagery will be publicly released shortly in Intergraph's
ImageStation Automatic Elevations - Extended.
The discussed results show that our SGM implementation can
collect info clouds from ADS and frame data, delivering a pro-
duct superior to LiDAR in geometric and spectral resolution. It
can be used instead of or in combination with LiDAR for many
applications; some of which have yet to be explored.
An important step to be addressed in the future is the merge of
info clouds. The final goal is to compute large, consistent info
cloud mosaics, which also form the basis for derived products
22
such as high-resolution DSMs and true ortho-image mosaics,
for both frame and ADS imagery.
7. REFERENCES
Gehrke, S., 2012. Combined Geometric/Radiometric Point
Cloud Matching for Shear Analysis. Int. Arch. Phot. & Rem.
Sens., Melbourne, Australia, Vol. 39, Part B1.
Gehrke, S., K. Morin, M. Downey, N. Boehrer, and T. Fuchs,
2010. Semi-Global Matching: An Alternative to LiDAR for
DSM Generation? Int. Arch. Phot. & Rem. Sens., Calgary, AB,
Vol. 38, Part Bl.
Gehrke, S., R. Uebbing, M. Downey, and K. Morin, 2011.
Creating and Using Very High Density Point Clouds Derived
from ADS Imagery. Proc. ASPRS Annual Conference, Milwau-
kee, WI.
Hirschmüller, H., 2005. Accurate and Efficient Stereo Proces-
sing by Semi-Global Matching and Mutual Information. Proc.
IEEE Conference on CVPR, New York, New York.
Hirschmüller, H., 2008. Stereo Processing by Semiglobal Mat-
ching and Mutual Information. IEEE Transactions on Pattern
Analysis and Machine Intelligence, Vol. 30, No. 2, pp. 328-341.
Jacobsen, K., 2011. Geometric Property of Large Format Digital
Camera DMC II 140. Photogrammetriue — Fernerkundung —
Geoinformation (PFG), Vol. 2011, No. 2, pp. 71-79.
Legat, K., 2012. Dense Image Matching — Initial Results. Euro-
SDR Workshop, Vienna, Austria, http://www.ipf.tuwien.ac.at/
wm/download/12 ws dsm eurosdr/16 Legat LinzExperiences.
pdf.
Sandau, R., Braunecker, B., Driescher, H., Eckardt, A., Hilbert,
S., Hutton, J., Kirchhofer, W., Lithopoulos, E., Reulke, R.,
Wicki, S., 2000. Design Principle of the LH Systems ADS40
Airborne Digital Sensor. /nt. Arch. Phot. & Rem. Sens., Amster-
dam, The Netherlands, Vol. 33, Part Bl, pp. 258-265.
Z/I Imaging, 2011. Camera Comparison. Brochure, ZI Imaging.
ABS
This
intei
lase
buil
the :
vari:
com
Thre
pert:
men
ing
reve
gam
in Oo
scan
Terr
intrc
accu
clos
lar i
tion
How
fore
ners
togr
the |
till t
cam
com
the
of tc
that
the :
poin
ima;
tion
lot t
the
once
cam
the :
togr
field
Now
