Figure 10: The extracted road network presented in CAD-based 
environment 
In order to evaluate the obtained results, the road network 
available in the test area was manually extracted to produce a 
road network reference map, which is shown in Figure 11. 
The comparison between Figures 10 and 11 showed that 
wherever a road pixel is extracted by the automatic procedure, 
its geometric position is exactly the same as manually produced 
road network, but the completeness of extracted road network is 
with doubt. Figure 12 shows those areas where the designed 
automatic procedure failed to detect road pixels, which are 
presented in boxes. 
4. CONCLUSIONS AND RECOMMENDATIONS 
Image processing techniques such as pattern recognition and 
object detection algorithms do not provide fully structured data 
for GIS environments. This is mainly due to the fact that the 
results of image processing algorithms are raster maps while 
GIS environments need structured vector based data as their 
input entities. Furthermore, it should be noticed that CAD based 
facilities could not be performed directly on source imaged 
because of the complicated nature of them, especially when 
dealing with high resolution ones. 
Therefore, the combination of image processing techniques and 
CAD facilities can be a good alternative to automate the process 
of data preparation and entering into GIS. 
A new approach was designed and implemented in this research, 
and a fully automatic procedure was designed and implemented 
to extract road centerlines from high resolution satellite images. 
Neural networks and digital image processing algorithms such 
as morphological operators were used to extract road centerline 
from satellite images and present it as an edited road raster map. 
Then obtained raster map was prepared and vectorized using the 
facilities provided by CAD-based environments. The final 
result is the structured vector based road network presented in 
CAD environment where it can be easily transformed to GIS. 
The comparison between obtained results and road network 
reference map proved acceptable geometric accuracy of 
designed procedure. 
The designed and implemented method in this research for 
automatic road extraction was mainly based on spectral and 
geometric properties of road networks. Knowledge-based road 
extraction methods, which are based on an appropriate road 
model, could be an efficient method for further study. In these 
methods, even when some parts of the road in input source 
image, road raster map or thinned road center line are missed, 
they could be recovered based on the road knowledge injected 
into the system via knowledge-based road model. 
REFERENCES 
GRUEN, A., AGOURIS, P, LI, H„ 1995. Linear Feature 
Extraction with Dynamic Programming and Globally Enforced 
Least Squares Matching. In: Gruen, A., Kuebler, O., Aqouris, P. 
(Eds.), Automatic Extraction of Man-Made Objects from Aerial 
and Space Images. Brikhauser, Basel. 83-94. 
GRUEN, A., LI, H., 1997. Semi-Automatic Linear Feature 
Extraction by Dynamic Programming and LSB-Snakes. 
Photogrammetric Engineering and Remote Sensing. 63(8): 
985-995. 
HARALICK, ROBERT M., AND LINDA G. SHAPIRO. 1992. 
Computer and Robot Vision. Volume I. Addison-Wesley, pp. 
28-48.