585 
THE APPLICATION OF NEURAL NETWORKS, IMAGE PROCESSING AND CAD- 
BASED ENVIRONMENTS FACILITIES IN AUTOMATIC ROAD EXTRACTION AND 
VECTORIZATION FROM HIGH RESOLUTION SATELLITE IMAGES 
F. Famood Ahmadi 3, *, M. J. Valadan Zoej a , H. Ebadi 3 , M. Mokhtarzade 3 
d K.N.Toosi University of Technology, Faculty of Geodesy and Geomatics Engineering, Tehran, Iran - 
farshid_famood@yahoo.com, -valadanzouj@kntu.ac.ir,- ebad)@kntu.ac.ir, m_mokhtarzade@yahoo.com 
Commission III 
KEY WORDS: Automatic, CAD, Image processing, High resolution images, Neural networks, Road extraction, Vectorization 
ABSTRACT: 
In this article a new procedure that was designed to extract road centerline from high resolution satellite images, is presented. The 
results (road Networks) are fully structured in vector formed in Computer Aided Design (CAD) based system that could be used in 
Geographical Information System (GIS) with minimum edit. The designed procedure is the combination of image processing 
algorithms and exploiting CAD-based facilities. In the first step, artificial neural networks are used to discriminate between road and 
non-road pixels. Then road centerlines are extracted using image processing algorithms such as morphological operators, and a road 
raster map is produced. Some cleaning algorithms were designed to reduce the existing noises and improve the obtained results. 
Finally, the edited raster map was vectorized using the CAD-based facilities. Obtained results showed that the structured vector 
based road centerlines are confirming when compared with road network in the reference map. 
1. INTRODUCTION 
Satellite and aerial images are the most important available data 
sources for map generation and updating of available maps. 
They provide accurate, easily accessible and reliable spatial 
information for Geographical Information Systems (GIS). The 
traditional manual methods of data capture from these images 
are expensive, laborious and time consuming and do not let full 
exploitation of available data in image archives. Nowadays 
when satellite images have highly improved in terms of spatial, 
spectral and temporal resolutions and Geomatics communities 
are overwhelmed by the sheer volume of collected images, the 
necessity of automation in feature extraction and map updating 
seems urgent. 
Roads as one of the most important man made objects are in 
high concern to be extracted (semi)automatically and many 
researches have been carried out in this area. Geometrically 
constrained template matching (Gruen et al., 1995; Vosselman 
and Knecht, 1995), active contours or snakes 
(Neuenschewander et al., 1995; Trinder and Li, 1995; Gruen 
and Li, 1997) and fuzzy set and morphological operators 
(Mohammadzadeh et al., 2006) are some of the semi-automatic 
methods for road extraction. 
Road extraction could be defined as the process of road 
identification and accurate localization in the image so that 
when the image to ground systems transformation is performed, 
the road network is truly presented in the object space. 
Automatic road extraction concentrates on automating all or 
some parts of this process to facilitate and expedite the road 
extraction task. 
In high resolution satellite images, roads could be regarded as 
elongated homogeneous regions that contrast from background 
with distinct spectral behavior. Based on this model, automatic 
road extraction from this kind of images can be categorized in 
three steps as road detection, road thinning and centerline 
extraction and finally vectorization of extracted road skeleton. 
Road detection is defined as the process of assigning a value to 
each pixel that can be used as a criterion to extinguish between 
road and background pixels. This process classifies the entire 
image into two different classes and has a major influence on 
the success of next stages. The segmented image, usually 
containing some unwanted and missed road pixels, is then 
introduced to some noise removal and other image processing 
algorithms to extract road centerline. Finally, the extracted road 
centerline is vectorized and transformed into CAD-based 
environments to be ready for GIS applications. 
In this research, a back propagation neural network with its 
different input parameters is proposed for road detection, which 
is described in section 2. This is followed by morphological 
thinning and other image processing algorithms for road 
centerline extraction accompanied by noise reduction and 
quality improvement techniques, as well as automatic 
vectorization, which is outlined in section 3. Conclusions and 
recommendations for further studies are presented in section 4. 
2. ROAD DETECTION USING ARTIFICIAL NEURAL 
NETWORKS 
Neural Networks are computational systems made up of simple 
processing units called neurons which are usually organized 
into layers with fully or partially connections. The main task 
associated with a neuron is to receive the activation values from 
its neighbors (the output of other neurons), compute an output 
based on its weighted input parameters and send that output to 
its neighbours. 
Corresponding author.