5
Model Generalization and Methods
for Effective Query Processing and Visualization
in a Web Service/Client Architecture
Marian de Vries and Peter van Oosterom
Delft University of Technology, Delft (The Netherlands)
5.1 Introduction
Generalization is a long-standing research subject in geo-science. How to derive
1:50,000 scale maps from 1:10,000 scale maps, for example, has been an issue also
in the pre-Internet age. But now, because of distributed geo-processing over the Web
(Web mapping and Internet GIS), research into automated generalization of geo-data
received a new impulse. One of the research issues is how to enable ‘zoom-level
dependent’ retrieval and visualization of geo-data in Web service/client contexts.
In a Web context, to be able to display the right amount of detail on a map in
a Web client is important for a number of reasons. First of all there is the ‘classic’
requirement of cartographic quality. Too much information will clutter the map image and make it unreadable and too little detail is not very useful either. Secondly,
especially because of the Web service/client context, there is the requirement of response time: the LOD of the geo-information that is displayed in the Web client will
influence both the amount of bytes transferred over the network and the visualization (rendering) time in the client. Generalization at the Web-server side, before the
information is sent to the client, has therefore two advantages: the user gets the LOD
that is needed at that zoom level, and response time as experienced by the user will
be reduced.
Generation and display geo-information at the right LOD can be tackled in different ways. One approach is to use multiscale/multirepresentation databases that are
created in an offline generalization process, and then switch to the appropriate representation for a certain zoom level when the user zooms in or out (e.g. [7]). Another
approach is to use on-the-fly generalization, where, for example, line simplification
is carried out in real time as part of the visualization process (in the Web client or by
a specialized generalization Web service, see [3]).
In this chapter we present a third approach, based on a variable scale data structure complex: the topological GAP structure (generalized area partitioning structure),
or tGAP structure for short. The purpose of this structure is to store the geometry only
once, at the original, most detailed, resolution. There is an offline generalization process, which builds a hierarchical, topological structure of faces (the GAP-face tree)
Model Generalization and Methods
for Effective Query Processing and Visualization
in a Web Service/Client Architecture
Marian de Vries and Peter van Oosterom
Delft University of Technology, Delft (The Netherlands)
5.1 Introduction
Generalization is a long-standing research subject in geo-science. How to derive
1:50,000 scale maps from 1:10,000 scale maps, for example, has been an issue also
in the pre-Internet age. But now, because of distributed geo-processing over the Web
(Web mapping and Internet GIS), research into automated generalization of geo-data
received a new impulse. One of the research issues is how to enable ‘zoom-level
dependent’ retrieval and visualization of geo-data in Web service/client contexts.
In a Web context, to be able to display the right amount of detail on a map in
a Web client is important for a number of reasons. First of all there is the ‘classic’
requirement of cartographic quality. Too much information will clutter the map image and make it unreadable and too little detail is not very useful either. Secondly,
especially because of the Web service/client context, there is the requirement of response time: the LOD of the geo-information that is displayed in the Web client will
influence both the amount of bytes transferred over the network and the visualization (rendering) time in the client. Generalization at the Web-server side, before the
information is sent to the client, has therefore two advantages: the user gets the LOD
that is needed at that zoom level, and response time as experienced by the user will
be reduced.
Generation and display geo-information at the right LOD can be tackled in different ways. One approach is to use multiscale/multirepresentation databases that are
created in an offline generalization process, and then switch to the appropriate representation for a certain zoom level when the user zooms in or out (e.g. [7]). Another
approach is to use on-the-fly generalization, where, for example, line simplification
is carried out in real time as part of the visualization process (in the Web client or by
a specialized generalization Web service, see [3]).
In this chapter we present a third approach, based on a variable scale data structure complex: the topological GAP structure (generalized area partitioning structure),
or tGAP structure for short. The purpose of this structure is to store the geometry only
once, at the original, most detailed, resolution. There is an offline generalization process, which builds a hierarchical, topological structure of faces (the GAP-face tree)