tial
lel
on,
ly,
ing
1est
of
lata
ure
ited
are
tial
ting
tial
ese
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
Road information can be 1-D (linear referencing systems), 2-D
(planar coordinates), 3-D (planar coordinates and height
information) and 4-D (time in case of dynamic objects)
depending on the usage. The road object is associated with three
different referencing systems being (X, Y, Z), (1,9), which
describes linear referencing system and (h.q), defining cross-
sectional road data, illustrated in Figure 1.
q
Figure 1: Reference Systems
Excluding the planar coordinates, other dimensions, which have
implicit references such as road identifier, name, stationing
value, gradient value, are considered as non-spatial information.
This approach concludes several problems during the system
maintenance, since road object is subject to change. Redundant
information storage and mismatch of information is common.
Since road objects can be static or dynamic, referenced to one,
two, and three or four dimensions, the traditional road data
models are insufficient to support data integration. In the
transportation context, three classes of GIS models are relevant,
being: i) field nodels, i) discrete models, and i) network
models (Goodchild, 1998). The network model is the commonly
used one among these, since many applications only require a
network model to represent data (Thill, 2000). The network data
models represent topologically connected linear entities that are
fixed in the continuous reference surface, inherently 2-D. In the
model, arcs and nodes themselves are primitives of the discrete
entity model, stored as geometric structures. Associating I-D
data with the 2-D planar model is a well-known problem, since
an attribute was described as a spatial (1-D) event occurring on
the network, however in many cases I-D data requires to be
associated with 3-D data.
In order to comprehend the user requirements and to diminish
the termined problems, a generic conceptual data model
needs to be designed after constitution of a criteria list.
Considering the peculiarity of spatial information, several
systems were analyzed (Vonderohe ct al.. 1997; Walter, 1997;
Vejdirektoratet, 1998; NWSIB, 1998, NCHRP, 1998: Gielsdorf,
1998; BMVBW, 1998; OKSTRA, 2000; Dueker and Butler,
2000; Sutton and Wymann, 2000; Adams et al., 2000; Portele,
2001; Koncz and Adams, 2002). According to these studies, a
model for this purpose should ensure: 1) conceptual
decomposition of topological, geometric and thematic
information, ii) referencing multi-dimensional road information
in 3-D and time, lil) transformation of datasets, iv) support for
multiple topological representation and various abstraction
levels, v) non-planar topological model, vi) history information,
and vii) incorporation of metadata, involving accuracy and
integrity constraints.
3. The Conceptual Data Model
The designed conceptual data model is constituted upon the
criteria list. During the establishment of external schema, a
progressive approach appropriate to the conceptual data
modeling requirements of an entire road administration was
reflected on. The data model was designed with four distinct
components: geometry, topology, road events and metadata.
The main approach of the proposed data model is abstraction
and decomposition of geometry, topology and non-spatial data.
The basic component of the proposed data model is geometry,
which the model is diverging from standard GIS models. In
most GIS, geometry plays a secondary role compared to
thematic data and conventionally used for graphical
visualization. Consequently, it is redundantly implemented. The
geometry component here is subdivided into three categories;
point geometry, linear geometry and area geometry. According
to the proposed conceptual data model, a complex object is an
aggregation of several simple objects. The point geometry is
defined in terms of a three-dimensional coordinate system,
including height information. In order to achieve data
integration, control of redundancy and optimization of data
maintenance, linear elements were mapped by means of datum
invariant parameters, which are either horizontal or vertical
planes. The planar linear elements have threc parameter types;
line, arc and clothoid. Area geometry may be either planar or
non-planar. By considering geometry to be the basic
component, many of the problems noted are avoided.
Topology, being non-planar and having two abstraction levels,
as modeled as a logical abstraction of geometry. Two
abstraction levels are required for the multi-scale
representations. The multi-scale representations result from
seeing the world from different abstraction levels, as well as
different points of view (Bedard et al., 2002). By possessing
topological information implicitly, temporal management
becomes easy. The main elements of topology were node and
link. According to the conceptual data model, Link is a logical
connection defined by two nodes. Topology elements are
adapted from traditional planar networks. Associations between
nodes and links were applied to both abstraction levels.
Additional associations were defined for mapping the
associations between two abstraction levels. Relationships
between abstractions are modeled as follows: ‘Node I’, being a
higher abstraction level may be composed of ‘Link II’ and
‘Node II". Between the first level topology element node (‘Node
I’) and the second level topology element node (*Node II) a 1:
0..* relationship is assigned, where ‘Node I’ may be composed
of many nodes (‘Node Il’) and every ‘Node II’ is assigned to
only one object ‘Node I’. The relationship between ‘Node I’ and
‘Link II’ is modeled as 0..1: 0..*, where ‘Node I’ may be
composed of second level links (‘Link Il) and a second level
link (*Link I) may be assigned to none or only one ‘Node I’.
Road junctions can be modeled using these associations. The
relationship between ‘Link I’ and ‘Link IP is modeled as 0..1:
0..*, where ‘Link I’ may be composed of second level link’s
(‘Link IP) and a second level link (‘Link II’) may be assigned to
‘Link I’. By means of defined abstraction levels. support for
multiple topological representation and various abstraction
levels were incorporated into the system. Additionally, defining
the geometry components in 3D allows implementing a non-
planar topological model. Implementation results can be found
in Demirel (2002).
1313
