International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004
Fig. 11 presents two examples (3 images in one row for each) to
show the limitation when our system is applied to HRS data. In
the figure, the VEC25 roads and extracted results are presented
as white and black lines respectively, while the Swissimage,
IKONOS and Quickbird orthoimages are shown from left to
right. In both examples, the roads are extracted from
Swissimage. The road shown in the first scene (first row) is not
extracted in the IKONOS image, while the road in the second
scene (second row) can not be extracted in the HRS data due to
haze. Clouds prohibit road extraction in the example of Fig. 12.
Figure 11. Examples showing limitations of our system applied
to HRS data. See text for
Figure 12. Example showing clouds in the Quickbird image
preventing road extraction. Left: Swissimage,
middle: IKONOS, right: Quickbird. The VEC25
road and extraction results are presented in white
and black lines respectively.
5. CONCLUSIONS
In this paper, we have reported the performance comparison of
the ATOMI road reconstruction system between aerial film
orthoimages of varying pixel size, ADS40 and HRS
orthoimages over two test sites in Switzerland, using accuracy,
completeness and correctness quantitative measures and visual
control. It is shown that about 9594 of roads in rural arcas are
correctly extracted using aerial film and ADS40 orthoimages
with 20-cm and 30-cm pixel size, respectively. With increasing
pixel size, the system performance deteriorates but to a much
less degree. However, even though the landcover of the two test
sites is largely different, our system achieved in both ca. 9094
completeness with 50cm aerial orthoimages. Thus, the general
conclusion is that the ATOMI system can reconstruct road
networks in rural areas using aerial orthoimages with maximum
pixel size of ca. 50-60 cm with a completeness and correctness
of 90%-95% and an accuracy of 0.4-0.7 m. The speed is
sufficient for operational production, while the result includes
both extracted and non-extracted (old) data resulting in a
complete network with the topology and attributes of the input
road database plus new derived attributes like road width. Using
manual on-screen digitising of road seed points, the method can
be extended to generation of a road database from scratch.
746
The test shows that the system performance is poor with the
HRS data, especially for 1-m IKONOS PSM imagery. Both
HRS can deliver submeter accuracy, however the problem lies
with the poor object definition and image quality. Only half of
the roads in the test sites are reconstructed, mainly higher-class
roads with larger width. The surface of the narrow roads (lower-
class roads) is usually blurred and the road edges are weak and
not clear in the HRS images, thus most of the lower-class roads
are not extracted. The test results show that the performance on
the 70-cm Quickbird data is considerably better than that on the
1-m IKONOS data, but still of lower completeness than the 60-
cm pixel size aerial orthoimage. Other extraction methods, not
requiring road widths (ribbons) of 3 pixels or more, may be
more appropriate for orthoimages with such pixel size.
Generation of Quickbird orthoimages with 60 cm, or
deployment of new HRS with 40 cm - 50 cm ground pixel size
(license for which the US government has already provided)
may pave the way for application of the current approach with
good completeness even for such imagery, if the imaging
conditions are favourable.
Our system can still be improved, for example, by better use of
the existing road vectors to bridge gaps. Post-control on
whether the solution conforms in shape and topology to road
construction and intersection principles needs to be completed.
The self-diagnosis and reliability measures derived for the
extraction results are not robust enough. Use of denser and
more accurate laser DSMs/DTMs and of the NIR channel of
digital sensors can be used for better quality results. Extension
of the method to areas with low buildings and forest borders
may be feasible. These and other aspects will be topics of future
research.
ACKNOWLEDGEMENTS
We acknowledge the financial support and the data provided for
this work and the project ATOMI by the Swiss Federal Office
of Topography and the NPOC, Bern. Canton Geneva provided
25-cm orthoimages, a laser DSM and other data in the Geneva
test site. Space Imaging USA provided the Rational Polynomial
Coefficients for the IKONOS images. We also thank Zhang Li,
Henri Eisenbeiss, Oliver Heller and Oliver Gut at ETH Zurich
for providing the orthoimages of IKONOS and Quickbird.
REFERENCES
Baltsavias, E., Zhang, C., 2003. Automated updating of road
databases from aerial imagery. Proc. Workshop "Data
quality in Earth Observation Techniques", ITC, Enschede,
The Netherlands, 21 November.
Eidenbenz, Ch., Kaeser, Ch., Baltsavias, E.P.. 2000. ATOMI —
Automated Reconstruction of Topographic Objects from
Aerial Images using Vectorized Map Information.
International Archives of Photogrammetry, Remote Sensing
and SIS, Vol. 33, Part B3/1, pp. 462-471.
Heipke, C., Mayer. H., Wiedemann, C., Jamet, O., 1998.
External evaluation of automatically extracted road axes.
Photogrammetrie Fernerkundung Geoinformation (2), 81-94.
Zhang, C., 2003a. Updating of cartographic road database by
image analysis. Ph.D. Thesis, Institute of Geodesy and
Photogrammetry, ETH Zurich, Switzerland, Report No. 79.
Zhang, C.. 2003b. Towards an operational system for
automated updating of road databases by integration of
imagery and geodata. ISPRS Journal of Photogrammetry and
Remote Sensing 58(3-4), 166-186.
UP]
KEY
ABS]
Since
parts
useful
image
photo
from :
currer
which
Or cot
are pr
High
Quick
especi
inacce
such
Castre
updat
areas
(Man
these
have
often
proce:
can pr
of im:
variet:
In tre
Weste
detail
over r
techni
and pr
count
rather
busine
satelli
data s
world
to dis
used t
countr
At Or
à proj
Quick
scale
larger,
the IS
