357
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
that have different resolution. Six classes were defined; and for
each class training data were collected and images were
classified. Then classified images were reclassed as two classes;
road and nonroad by merging classes other than road and results
were compared. Although the classified data of high resolution
images produced better results, several problems experienced;
When the pavement of road surface changes (e.g. from asphalt
to concrete) the contrast changes so the same road could be
assigned different cluster label.
Due to very high resolution, vehicle, overpass, and other objects
on a road could cause misclassification for that particular part.
Other objects such as house roofs, which have same spectral
property as roads, are misclassified into road clusters.
Figure 4: Classified images with two classes.
3. CONCLUSION
• Many objects which can be recognized from high
resolution images can not be detected distinctly in a low
resolution image.
• In a low resolution image roads can be extracted by
using basic methods but the results have low accuracy.
• In a high resolution image many detail information
about roads can be obtained and the roads can be
detected more accurately. However complex methods
are required for this process.
• Finally, the accuracy of road extraction from high
resolution images is more than that one from low
resolution images.
REFERENCES
Baltsavias, E., Gruen, A., van Gool, L. (Eds.), 2001.
“Automatic Extraction of Man-Made Objects from Aerial and
Space Images (III). Balkema, Lisse, The Netherlands”.
Fortier, M.F., Ziou, D., Armenakis, C. and Wang, S., 1999.
“Survey of Work on Road Extraction in Aerial and Satellite
Images”. Center for Topographic Information Geomatics,
Ontario, Canada. Technical Report No. 241.
Gruen, A., and H. Li, 1997. “Semiautomatic linear feature
extraction by dynamic programming and LSB-snakes”,
Photogrammetric Engineering & Remote Sensing, 63(8):985-
995.
Gruen, A., Baltsavias, E., Henricsson, O. (Eds.), 1997.
“Automatic Extraction of Man-Made Objects from Aerial and
Space Images (II)”. Birkha'user Verlag, Basel.
Koike, K., Nagano, S. And Ohmi, M., 1995, “Lineament
Analysis of Satellite Images Using A Segment Tracing
Algorithm (STA)”, Computers and Geosciences, Vol. 21, No. 9,
1091-1104.
Ruisheng Wang ve Yun Zhang, 2003.“Extraction of urban road
network using Quickbird Pansharped Multispectral and
Panchromatic Imagery by Performing Edge-Aided Post-
Classification”
http://www.geoict.net/Resources/Publications/RW_UNB_GRS
S_2003.pdf
Wiedemann, C., Hinz, S., 1999. “Automatic extraction and
evaluation of road networks from satellite imagery”.
International Archives of Photogrammetry and Remote Sensing,
vol. 32, Part 3-2W5, pp. 95-100.
Xiangyun Hu C. Vincent TaoYong Hu, 2004 “Automatic road
extraction from dense urban area by integrated processing of
high resolution imagery and lidar data”
http://www.isprs.org/istanbul2004/comm3/papers/288.pdf
Zheng, Jeffrey, Liu, Gang-Jun, and Coppa, Isabel. 1998.
“Linear Geo-spatial feature extraction from remotely sensed
digital imagery using conjugate transformation”. 9th Australian
Remote Sensing and Photogrammetry Conference, vol. 1, pg
6401.