3. Istanbul 2004 
t road segment 
olygon with the 
d road segment 
"ht line segment 
road segments 
ther analyses of 
in evidence that 
d. For example, 
fore, everything 
action problem 
atic completion 
Ss 
dology for road 
image data are 
oriate for road 
solution images 
1 and another of 
  
olution image 
dium-resolution 
ads manifest as 
image and the 
o construct the 
id on the input 
h of 6 pixels is 
only main road 
tained can be 
ptions are short 
ch should be 
ded to increase 
ed due to the 
oads. For these 
objects can be 
road network, 
acted. Specific 
road segments 
    
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
connection are needed to be developed to make possible the 
extraction of the complete road network. Anyway, the method 
allows long road segments to be extracted, facilitating the 
posterior automatic completion of the road network. In terms of 
completeness, about 80% of the road network is extracted. 
  
Figure 6. Results obtained for the high-resolution image 
The second experiment (figure 6) is carried out with a high- 
resolution image (498 x 535 pixels), in which the roads 
manifest as wide ribbons with 40-pixel width. Figure 6 shows 
that the results reflect those theoretically expected. In fact, only 
road parts perturbed by illumination posts casting on the road, 
and the road crossing, are not extracted as those parts are 
incompatible with any road objects. As already mentioned in 
previous experiment, these fails are needed to be treated by 
specific strategies embody another types of road knowledge, 
like ones based on context and scale space (Baumgartner et al., 
1999). Note that the vegetation edges adjacent to road edges do 
not cause false positives, showing the robustness of the method 
in these situations. As in previous experiment, the completeness 
is also about 80%. 
4. CONCLUSIONS AND FUTURE PERSPECTIVES 
This paper presented an automatic method for the road segment 
extraction from medium- and high- resolution images of rural 
scenes. The innovation in the proposed methodology is the way 
the road objects and the connection rules between them are 
defined. 
With purpose of evaluating the method's potential in extracting 
road segments, two experiments were conducted using two test 
images, being one of high-resolution and another of medium- 
resolution. In all cases the results obtained can be considered 
satisfactory as they are in accordance to ones theoretically 
expected. Some little disconnections are expected, as the edge 
detection is sensible to the irregularities along the road margins. 
As a result, no road objects can be constructed for such road 
parts, given rise to the missing road segments. Also due to 
incompatibility with any road objects, road crossings were not 
extracted by the proposed methodology. Despite these 
theoretical expected fails, the methodology was able to extract 
about 80% of the road network. The automatic road network 
completion methodologies will be the focus of our future 
researches, whose basic input will be the road segments 
extracted by the proposed methodology. 
ACKNOWLEDGEMENTS 
This work has been supported by FAPESP (Research 
Foundation of the State of Sao Paulo, Brazil) and CNPq 
(National Council of Research and development, Brazil). 
REFERENCES 
Bajcsy, R., Tavakoli, M., 1976. Computer recognition of roads 
from satellite pictures. /EEE Transactions on Systems, Man, 
and Cybernetics, 6 (9), pp. 76-84. 
Baumgartner, A., Steger, C., Mayer, H., Eckstein, W. Ebner, 
H., 1999. Automatic road extraction based on multi-scale, 
grouping, and context. Photogrammetric Engineering and 
Remote Sensing, 66 (7), pp. 777-785. 
Doucette, P., Agouris, P., Stefanidis, A., Musavi, M., 2001. 
Self-organized clustering for road extraction in classified 
imagery. ISPRS Journal of Photogrammetry and Remote 
Sensing, 55, pp. 347-358. 
Doucette, P., Agouris, P., Musavi, M., Stefanidis, A., 2000. 
Road centerline vectorization by self-organized mapping. In: 
International Archives of the Photogrammetry, Remote Sensing 
and Spatial Science, 33, Part B3, Amsterdam, pp. 246-253. 
Quam, L. H., 1978. Road tracking and anomaly detection in 
aerial imagery. In: Image Understanding Workshop, pp. 51-55. 
Wiedmann, C., Hinz, S., 1999. Automatic extraction and 
evaluation of road networks from satellite imagery. In: 
International Archives of the Photogrammetry, Remote Sensing 
and Spatial Science, Munich, 33, pp. 95-100.