In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Voi. XXXVIII, Part 7B 
458 
aperture radar (SAR) imagery. Clode et al. (2004), who used 
data only laser scanner for automated extraction of roads. 
Maillard and Cavaye (1989) developed methodologies for the 
extraction of roads using only the multispectral images. Zhang 
and Murai (1999); Mohammadzadeh et. al. (2004) introduces an 
extraction based on mathematical morphology. Among the 
existing solutions can highlight the use of Mathematical 
Morphology, which includes the study of topological and 
structural properties of objects from their images. (Gonzales; 
Woods, 2000). 
Given the need for an analysis of the potential of ALOS images, 
development of methodologies for the roads extraction and the 
study of problems involved in the process, this paper present a 
method to road extraction with ALOS images through the use of 
mathematical morphology. The paper have the following 
specific objectives: Analyze the potential of use ALOS images 
for the roads extraction and evaluate the use of mathematical 
morphology in this case. 
2. THE ALOS SATELLITE 
The ALOS is a Japanese satellite that was launched by the 
Japanese Space Agency (JAXA) on January 24, 2006, becoming 
operational on October 20, 2006. This was launched by the 
rocket H-IIA from Tanegashima Space Center, Japan. His 
original name in Japanese language is "DAICHI. 
The ALOS satellite carries onboard 3 sensors: 
AVNIR-2: Advanced Visible and Near Infrared Radiometer - 
type 2; 
PRISM: Panchromatic Remote-sensing Instrument for Stereo 
Mapping; 
PALSAR: Phased Array L-band Synthetic Aperture Radar. 
In this work it is worth highlighting the AVNIR-2 and PRISM. 
The AVNIR-2 is an optical sensor with 4 spectral bands, 3 
bands of the visible and 1 near infrared band with a spatial 
resolution of 10 meters. The PRISM sensor operates in the 
visible light, with a panchromatic band and spatial resolution of 
2.5 m. This is a set of 3 independent imaging telescopes that 
allow scenes nadir, leaning forward and leaning back. This 
system makes possible the acquisition of stereoscopic images 
along the path. 
3. MATHEMATICAL MORPHOLOGY 
The mathematical morphology has been widely used in digital 
image processing and focuses on the area that studies the 
geometric properties of objects in the images. This allows the 
extraction of image components that are useful in the 
representation and description of the shape of a region, such as 
borders and skeletons (Gonzales; Woods, 2000). The extraction 
of elements present in an image is accomplished with the help 
of a suitable structural element. The structural elements are 
matrices responsible for the removal or addition of labeled 
pixels in the image, which depend on their size and shape, 
usually defined by the User, according to the area to be applied. 
In this paper the mathematical morphology adopted is binary, 
thus only binary morphological operators were used. The basic 
operations of morphology are erosion and dilation, at the first 
the pixels that do not conform to a given pattern are deleted 
from the image and at the second a small area related to a pixel 
is changed to a given pattern. These operations are the basis for 
most of the operations used in mathematical morphology, in 
other words, they are combinations of these such as opening, 
closing, skeletonization, among others. 
The dilation is a morphological operation that combines two 
sets using the vector addition of elements of sets. Its symbol is 
®, the result as the name suggests is a dilated image, like this 
the effect of the dilation on an image is the growth or expansion 
of the object. These objects refer to the pixels whose gray level 
is greater than zero in relation with the background. The 
dilation can also be understood as the union of translations of A 
by elements of B. The Dilation of a set A by B denoted by A ® 
B, is given as: 
A ® B = (x | (B) xnA/0) (1) 
where A represents the image being operated on and B is called 
structuring element and its composition defines the dilation, so 
the dilation expands an image. 
Therefore, the dilation of A by B is the set of all x 
displacements such that B and A overlap in at least one nonzero 
element. 
Unlike the dilation, the erosion reduces the object present in the 
image against the background. It is a morphological operation 
that combines two sets using vector subtraction of elements of 
sets, its symbol is ©. The erosion of A by the structuring 
element B, denoted by A © B, is given as 
A©B = (x|(B)xcA) (2) 
The erosion of A by B is the set of all points x such that B, 
when translated by x, be contained in A. 
The result of successive erosions and dilations allow the 
elimination of specific details of the image, smaller than the 
structuring element without distortion of the features not 
deleted. The effect of re-application is no longer to modify the 
previously transformed result. 
The opening operator is used to remove parts of objects or even 
objects smaller than the structuring element. Thus, the opening 
operator can eliminate noise due to the erosion operator that is 
applied initially. The opening of a set A by B, denoted by A 0 B 
is given by equation 3. 
A ° B = (A0B) ® B (3) 
where A represents the image being operated on and B is called 
structuring element. 
The closing operator tends to join "islands" which the distance 
between them is less than the structuring element and closing 
holes smaller than this element. Being the same set A and a 
structuring element B, the closing of A by B, denoted by A • B, 
is given by equation 4. 
A*B = (A®B)©B (4) 
Soon, these operated jointly applied enable the formation of the 
most compact and at the same time, eliminate regions very small 
or thin.