Me
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
Candidate polygon
Base polygon
Figure 3. Examples of possible combinations between the base
and candidate polygons
Figure 3 shows the combination of polygons representing the
edges of a same road. One of polygons being combined is
labelled base and another one is labelled candidate. As also
shown in figure 3, the base polygon can be compatible with
more than one candidate polygon. This means that a base
polygon needs to be combined with candidate polygons until
the whole base polygon is combined. Assuming that every
combination gives true results, parts of or whole polygons (base
or candidate) are removed from the search space after they are
combined.
Road
centerline
Linking between
road segments
(b)
Figure 4. Extraction of road segments. (a) Extraction of road
objects; and (b) Connection of the road objects
The application of methodology described above to the
illustrative example of figure 3 would allow the result shown in
figure 4 to be obtained.
Supposing that the base polygon is combined from the right to
the left (figure 4(a)), two road objects are constructed by
combining the base polygon with the first candidate polygon.
The connection between these two road objects generate a first
road segment (figure 4(b)), which in turn can be decomposed
into three consecutive quadrilaterals. The vertices of these
quadrilaterals allow the definition of a short road segment
centreline. Now the combination of the base polygon with the
second candidate polygon generates the second road segment
constituted by only one quadrilateral. The straight line segment
of the base polygon not integrating to both road segments
extracted, may be a useful information for further analyses of
the reasons for the missing road segment. It is an evidence that
both extracted road segments are anyway related. For example,
there could be a "T" or "Y" road crossing. Therefore, everything
providing information on road segment extraction problem
should be preserved for further use in automatic completion
strategies of the road network.
3. EXPERIMENTAL RESULTS
In order to evaluate the potential of the methodology for road
segment extraction, two experiments with real image data are
carried out. As this methodology is appropriate for road
segment extraction from medium- and high- resolution images
of rural scenes, one image of medium-resolution and another of
high-resolution are used.
Figure 5. Result obtained with the medium-resolution image
The first experiment is carried out with a medium-resolution
image (500 x 500 pixels), in which the main roads manifest as
ribbons with 6-pixel width. Figure 5 shows this image and the
results overlaid on it. Parts of polygons used to construct the
road segments and their centrelines are overlaid on the input
image. In this experiment, an average road width of 6 pixels is
used to set up the road objects, meaning that only main road
segments can be extracted. The results obtained can be
considered close to the expected one. Two exceptions are short
road segments pointed out by arrows, which should be
extracted. Thus, new control strategies are needed to increase
the method efficiency. Other fails are expected due to the
missing edges and other anomalies along both roads. For these
places and especially for road crossings no road objects can be
constructed and, as a result, a fragmented road network,
composed by isolate road segments, is extracted. Specific
strategies for road crossing reconstruction and road segments
Internat
connecti
extractic
allows
posterio
complet
Fig
The sec
resoluti
manifes
that the
road pa
and the
incomp
previou
specific
like ong
1999). |
not cau
in these
is also :
4, |
This pa
extracti
scenes.
the roa
defined
With p
road se
images
resolut:
satisfac
expect:
detecti
Asar
parts, ;
incom
extract
theoret
