2004
/
~~,
nage
» (b)
| (d)
road
(g) (h)
= RES [Wel STE
(i)
Figure 6. (g) applying second trivial opening, (h) effect of
closing, (i) extracted road center line, and (j)
extracted road centerline in blue which is
superimposed on the original image
3. CONCLUSIONS
In this paper, an approach to detect road network from high
resolution image using combination of a developed fuzzy
system and mathematical morphology is proposed. In the fuzzy
stage, only mean value and standard deviation of road is
enough to classify the input image. The methodology has high
performance for hyper spectral images that different image
bands can be easily inserted or removed. Also it is tested that
the mentioned fuzzy approach is much faster than maximum
likelihood classification. Another advantage of this fuzzy
classification method is its ease in introducing on artificial
neural networks.
The algorithm in the mathematical morphology stage is based
on the assumption that road network forms an elongated area
which can be extracted as the connected components with
certain criteria. Trivial opening preserves the whole road
network and filter out the noises. Granulometry analysis was
performed with trivial opening to provide size information of
objects in the image. The results show that this approach
provides sufficient information from successive steps for
automatic road extraction and has satisfactory results for
updating of road databases and change detection issues.
4 REFERENCE
Agouris, P., Gyftakis, S., and Stefanidis, A., 1998. Using a
fuzzy supervisor for object extraction within an integrated
geospatial environment. In: International Archives of
Photogrammetry and Remote Sensing, Vol. 32, Part 11I/1, pp.
191-195.
Buckner, A, 1998. Model based road extraction for the
registration and interpretation of remote sensing data. In:
International Archives of Photogrammetry and Remote Sensing,
Stuttgart, Germany, Vol. 32, Part 4/1, pp. 85-90.
767
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
Gruen, A. and Li, H., 1995. Semi-automatic road extraction by
dynamic programming, ZSPRS Journal of Photogrammetry and
Remote Sensing, 50(4), pp. 11-20.
Konecny, G., and Schiewe, J., 1996. Mapping from digital
image data with specific reference to MOMS 02. ISPRS
Journal of photogrammetry and remote sensing, 51, pp. 173-
181.
Mayer, H., Laptev, I., and Baumgartner, A., 1997. Automatic
extraction based on Multi-Scale modeling, context and snakes.
In: /nternational Archives of Photogrammetry and Remote
Sensing, Vol. 32, Part 3-2W3, pp. 106-113.
Melgani, F., Hashemy, B., and Taha, S.M.R., 2000. An explicit
fuzzy supervised classification method for multispectral remote
sensing images. IEEE Transactions on Geoscience and Remote
Sensing, 38(1) 287-295.
Pigeon, L., Solaiman, B., and Toutin, T., 2001. Linear
planimetric feature domains modelling for multi-sensors fusion
in remote sensing. In: Proceedings of SPIE AeroSense -
International Symposium on Aerospace/Defence Sensing,
Simulation, and Controls, Orlando, Vol. 4051, 8 p.
Vincent, L., 1993. Morphological greyscale reconstruction in
image analysis: Applications and efficient algorithms. JEEE
Transaction on Image Processing, 2(2): 176-201.
Zhang, C., Murai, S. and Baltsavias, E., 1999. Road network
detection by mathematical morphology. In: Proceedings of
ISPRS Workshop on 3D Geospatial Data Production: Meeting
Application Requirements, Paris, France, Pages: 185-200.
