68
Michela Bertolotto
transmission and reconstruction are very efficient, as they can be performed simply
by means of a single pass through each color band of the image.
The purpose of raster data exchanging over the Internet is very often only
visualization. In this case, techniques for progressive raster transmission are usually
suitable and efficient. Even when measurements need be performed on the images
(e.g. for photogrammetric purposes), common compression mechanisms (such as
JPEG) have proved to be effective [30]. However, some applications involve direct
object handling and manipulation. In this case a vector representation is required and
preservation of topological and metric properties is a major issue. This is discussed
in the following sections.
4.2.2 Progressive Transmission of Meshes
Part of the spatial data in vector format is represented by means of triangular meshes.
For example, triangulations are used for digital terrain modeling and for real objects
surface reconstruction and rendering [1, 16].
As discussed in Chap. 3, several compression methods for triangular meshes
have been defined in the literature (see also [15] for a survey). These methods are
either based on optimal point decimation techniques or they exploit combinatorial
properties of triangulations for efficient encoding. In this section we cite some that
have been applied for progressive transmission. Among them, the progressive mesh
scheme proposed by Hoppe [27] represents a fundamental milestone. However, this
method does not have a satisfactory performance as, like many other traditional
methods [19, 53], it is topology preserving. This is important for data consistency
but it limits the level of simplification. More recent methods achieve higher compression by slightly modifying the topology of the input mesh [1].
Juenger and Snoeyink [28, 29] defined a parallel point decimation technique for
triangulated irregular network (TIN) simplification that allows progressive transmission and rendering of TIN data. Finally, De Floriani and Puppo [15] proposed
a general framework for multiresolution hierarchical mesh representation. Such a
model has a high storage cost. A more efficient encoding structure has been described in [16] to facilitate progressive transmission. Although these progressive
transmission techniques are effective and they have already produced several prototypes [1, 28], they can only be applied to data represented in the form of triangular
meshes.
4.3 Progressive Vector Transmission for Web Mapping
The increasing abundance of geographical map data sets on the Internet has recently
generated a lot of interest in downloading such data sets for public and commercial
use. Although several systems have been developed, Web mapping still suffers from
slow transmission and information overload. Indeed, vector maps usually occupy
very large data sets due to the inherent complexity of their geometry and topology.
Furthermore, they are often represented at high levels of detail (LOD).
Michela Bertolotto
transmission and reconstruction are very efficient, as they can be performed simply
by means of a single pass through each color band of the image.
The purpose of raster data exchanging over the Internet is very often only
visualization. In this case, techniques for progressive raster transmission are usually
suitable and efficient. Even when measurements need be performed on the images
(e.g. for photogrammetric purposes), common compression mechanisms (such as
JPEG) have proved to be effective [30]. However, some applications involve direct
object handling and manipulation. In this case a vector representation is required and
preservation of topological and metric properties is a major issue. This is discussed
in the following sections.
4.2.2 Progressive Transmission of Meshes
Part of the spatial data in vector format is represented by means of triangular meshes.
For example, triangulations are used for digital terrain modeling and for real objects
surface reconstruction and rendering [1, 16].
As discussed in Chap. 3, several compression methods for triangular meshes
have been defined in the literature (see also [15] for a survey). These methods are
either based on optimal point decimation techniques or they exploit combinatorial
properties of triangulations for efficient encoding. In this section we cite some that
have been applied for progressive transmission. Among them, the progressive mesh
scheme proposed by Hoppe [27] represents a fundamental milestone. However, this
method does not have a satisfactory performance as, like many other traditional
methods [19, 53], it is topology preserving. This is important for data consistency
but it limits the level of simplification. More recent methods achieve higher compression by slightly modifying the topology of the input mesh [1].
Juenger and Snoeyink [28, 29] defined a parallel point decimation technique for
triangulated irregular network (TIN) simplification that allows progressive transmission and rendering of TIN data. Finally, De Floriani and Puppo [15] proposed
a general framework for multiresolution hierarchical mesh representation. Such a
model has a high storage cost. A more efficient encoding structure has been described in [16] to facilitate progressive transmission. Although these progressive
transmission techniques are effective and they have already produced several prototypes [1, 28], they can only be applied to data represented in the form of triangular
meshes.
4.3 Progressive Vector Transmission for Web Mapping
The increasing abundance of geographical map data sets on the Internet has recently
generated a lot of interest in downloading such data sets for public and commercial
use. Although several systems have been developed, Web mapping still suffers from
slow transmission and information overload. Indeed, vector maps usually occupy
very large data sets due to the inherent complexity of their geometry and topology.
Furthermore, they are often represented at high levels of detail (LOD).