225
How Do You Model a Network for Geospatial Technology?
How Do You Model a Network
for Geospatial Technology?
Back in Chapter 5, we discussed how real-world items are modeled or represented using GIS. Any type of network is going to involve connections between locations, whether streams, power lines, or roads. Thus (to use Esri
terminology), in GIS a network in its most basic form is represented by a series
of junctions (point locations or nodes) that are connected to each other by
a series of edges (lines or links). For instance, in a road network, junctions
might be the starting and ending points of a road or the road intersections,
while the edge would be the line representing the road itself. When designing a road network, keep in mind that there may be many types of edges and
junctions to represent. For example, a city’s road network would have edges
that represent streets, highways, railroads, light-rail systems, subway lines, or
walking paths, while junctions may represent not only the starting and ending of streets, but also highway entrances and exits, freeway overpasses and
underpasses, subway stops, or rail terminals.
When dealing with all these different types of edges and junctions,
the connectivity of the network in GIS is essential when modeling it. With
proper connectivity, all junctions and edges should properly connect to one
another, while things that should not connect, do not connect. For example,
if a freeway crosses over a road via an overpass, the network connectivity
should not show that as a valid intersection allowing the street to turn onto
the freeway at that junction. If your vehicle navigation system leads you to
this point, then instructs you to “turn right onto the highway,” it’s impossible for you to do so, but the device thinks you should be able to because
of how the network data is set up. In the same way, a railroad line may intersect with a street, but the network should not have a connection showing
that the street could continue along the rail line. If this kind of connection
was built into the data, you could conceivably be routed to turn onto the
railroad line and continue on it toward your destination. It sounds silly to
think of driving your car on the railroad tracks, but due to incorrect network
data, this line would simply represent the next road to take to get to your
destination.
Thinking along these lines, other features of a road network must also
be included in the model. For instance, some streets may be one way, or
some junctions may not allow left-hand turns, or U-turns may not be permitted. These types of features need to be properly modeled for the system
to be an accurate, realistic model of the road network. Although you may
be able to see the “one way” street sign when you’re driving, if that feature
has not been properly set up in the network, the system would have no way
of knowing not to try to route cars in both directions along the street. As
discussed previously, an overpass or underpass should not show up as being
connected to the road network (to be a viable option for a turn—if a device
network a series of
junctions and edges
connected together for
modeling concepts such
as streets.
junction a term used
for the nodes (or places
where edges come
together) in a network.
edge a term used
for the linkages of a
network.
connectivity the
linkages between
edges and junctions of
a network.
Précédent

- 245/522

Suivant