1 Spatial Data on the Web: Issues and Challenges
3
1.1 Models for Representing Spatial Semistructured,
Multiresolution, and Multiscale Data
Data models and data structures for representing and managing spatial data, especially in geographical applications, have been well studied in the past two decades in
several research areas and with different aims.
In the area of spatial databases, much effort has been devoted to both the definition of a data model (and query language), to be adopted by database systems
in order to handle spatial data and integrate them with traditional information, and
the study of spatial data structures, access methods, and algorithms for increasing
the performance of query processing when spatial properties are considered. In this
area, spatial data means mainly vector data, i.e. the representation of the position,
shape, and extension of an object made by means of a geometry, which is described
by a finite set of points embedded in a reference space. Usually, for geographical data, only the linear geometry is adopted and the object shape is built up by
drawing open or closed polylines. Recently, ISO has also devoted a complete series
of standards to geographical data. In particular, the specification of a vector-based
geometry data model is given in the standard 19107 (called “Spatial Schema”) [9]
and an object-oriented conceptual model for the design of spatial databases using
the Spatial Schema as geometry model is given in the standard 19109 (called “Rules
for Application Schemas”) [10]. Moreover, the integration of temporal and spatial
dimensions has enriched the data models of constructs for the representation and
management of “moving objects.” The reader can find many papers and books about
data models, data structures, and access methods for spatial data; here we cite only
some of them [13, 19, 20, 22].
Besides vector data, other types of spatial information are quite relevant for
geographical applications: (1) remote sensing images and any other image of the
Earth’s surface; (2) grids of cells containing the measures of some physical
parameters surveyed on the territory (they are often used in many environmental
studies); (3) digital terrain models, describing, for example, terrain and geological
information, like the well-known digital elevation models (DEMs) and triangulated
irregular networks (TINs).
The first two types of the above listed spatial information are usually denoted
using the general term raster data and have been well studied also in the research
areas of image databases and image processing. Terrain representations based on
DEMs and TINs have been investigated by researchers in the area of geometric data
structures and algorithms.
The fast development of Web applications is providing to Internet users a huge
amount of spatial information represented in one of the above listed types. However, such a distributed environment poses new problems with respect to spatial data
modeling, in particular:
• Multiresolution. Each available data set can have a different resolution level,
and this level may be very far from the resolution level the user is interested in.
Therefore, it could be very useful for spatial data providers to be able to generate
3
1.1 Models for Representing Spatial Semistructured,
Multiresolution, and Multiscale Data
Data models and data structures for representing and managing spatial data, especially in geographical applications, have been well studied in the past two decades in
several research areas and with different aims.
In the area of spatial databases, much effort has been devoted to both the definition of a data model (and query language), to be adopted by database systems
in order to handle spatial data and integrate them with traditional information, and
the study of spatial data structures, access methods, and algorithms for increasing
the performance of query processing when spatial properties are considered. In this
area, spatial data means mainly vector data, i.e. the representation of the position,
shape, and extension of an object made by means of a geometry, which is described
by a finite set of points embedded in a reference space. Usually, for geographical data, only the linear geometry is adopted and the object shape is built up by
drawing open or closed polylines. Recently, ISO has also devoted a complete series
of standards to geographical data. In particular, the specification of a vector-based
geometry data model is given in the standard 19107 (called “Spatial Schema”) [9]
and an object-oriented conceptual model for the design of spatial databases using
the Spatial Schema as geometry model is given in the standard 19109 (called “Rules
for Application Schemas”) [10]. Moreover, the integration of temporal and spatial
dimensions has enriched the data models of constructs for the representation and
management of “moving objects.” The reader can find many papers and books about
data models, data structures, and access methods for spatial data; here we cite only
some of them [13, 19, 20, 22].
Besides vector data, other types of spatial information are quite relevant for
geographical applications: (1) remote sensing images and any other image of the
Earth’s surface; (2) grids of cells containing the measures of some physical
parameters surveyed on the territory (they are often used in many environmental
studies); (3) digital terrain models, describing, for example, terrain and geological
information, like the well-known digital elevation models (DEMs) and triangulated
irregular networks (TINs).
The first two types of the above listed spatial information are usually denoted
using the general term raster data and have been well studied also in the research
areas of image databases and image processing. Terrain representations based on
DEMs and TINs have been investigated by researchers in the area of geometric data
structures and algorithms.
The fast development of Web applications is providing to Internet users a huge
amount of spatial information represented in one of the above listed types. However, such a distributed environment poses new problems with respect to spatial data
modeling, in particular:
• Multiresolution. Each available data set can have a different resolution level,
and this level may be very far from the resolution level the user is interested in.
Therefore, it could be very useful for spatial data providers to be able to generate