Automated generalization of national topographic data within a Web environment

1
Automated generalization of national topographic
data within a Web environment

• Theodor Foerster
• NLS, Finland
• 27 September 2007

2
Overview

• Research introduction
• Generalization operator classification
• Web Processing Services (?)
• Discussion

3
Research context

• PhD study is embedded within DURP
• Physical planning
• Has been detached from topographic data
• Requires a base map for orientation and
  localization for different plans and users
• Physical plans have to be web-accessible
• DURP provides motivation use case!

4
Research components

• Focus of PhD study

Web
DURP
Topographic data
User
Automated generalization
5
Main requirements

• Produce generalized maps
• Cartographic generalization
• Automated generalization processing on the web
• Ensure meaningful generalization processing by
  semantics

6
Research statements

• The physical plan as a topological framework
• The web as a generalization processing environment

7
Conceptual architecture
Reasoner for generalization
Web Client
Generalization System
Physical plan data
Topographic data
8
Conceptual architecture

• The physical plan as a topological framework

Reasoner for generalization
Web Client
Generalization System
Physical plan data
Topographic data
9
Conceptual architecture

• The web as a generalization processing environment

Reasoner for generalization
Web Client
Generalization System
Physical plan data
Topographic data
10
Implementation architecture
Catalog
Stores queries
Reasons over
link
Reasoner
queries
getMap() SLD (incl. user profile)
WPS for generalization
executes
Forwards request
Clarity
WMS
Web Map Client
Takes into account
generalizes
getMap()
Topographic data WFS
Plan data WFS
Plan data WMS
visualizes
11
Main aspects

• User profiles drive the generalization process
• Generalization operators embedded as Web Services
• Semantics incorporated by reasoner
• Incorporating (WPS) profiles
• Incorporating more sophisticated semantic
  measures (semantic web)
• Classification of generalization operators

12
Generalization operator classification

• Based on

13
Overview

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

14
Introduction

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Generalization
• The adaptation of geo-data according to user and
  use requirements
• Model generalization (? data)
• Cartographic generalization (? maps)
• Generalization process consists of different
  generalization operators

15
Generalization operators

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Provide an abstract description of generalization
  functionality
• Important to abstract human knowledge
• Allows comparing generalization functionality
• An operator is implemented by different algorithms

16
Motivation

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Different classifications are available
• Do not cover all aspects
• Project-driven
• Not formal
• Classification is important
• For enabling automated web-based generalization
  processing
• Transparency for evaluation
• Knowledge exchange
• Producing on-the-fly base maps for different
  scenarios on the Web
• RGI 002 DURP ondergronden project
  (www.durpondergronden.nl)

17
Gruenreich model

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

Reality
Object generalization
Model generalization
Primary model
Cartographic generalization
Cartographic model
Inspired by Gruenreich (1992), ATKIS - a
topographic information system as a basis for GIS
and digital cartography in Germany
18
Classification framework

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

19
Model vs. cartographic generalization operators

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Model generalization operators can be defined on
  feature type level
• Cartographic generalization operators are defined
  on cartographic feature instance level
• Both are applied on instance level!

20
Operator classification

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Identified operators are based on literature

21
Model generalization operators

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Class Selection
• Reclassification
• Collapse
• Combine
• Simplification
• Amalgamation

22
Collapse Combine

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Changing the geometry type
• Collapse
• Decrease dimensionality
• Example Collapse road polygons to road lines
• Combine
• Increase dimensionality
• Goes along with reclassification
• Example Combine leisure sites (modeled as
  points) to touristic attraction (modeled as
  polygon)

23
Simplification Amalgamation

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Simplification
• Mainly used for data reduction
• Deletes aspects of the feature
• Not that invasive
• Amalgamation
• Amalgamating a group of features to a new feature
  of same geometry type
• Goes along (mostly) with reclassification
• Example Merging adjacent polygons of different
  types of forest to a new feature of type forest.

24
Role of symbolization

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Exclude symbolization from the cartographic
  generalization process
• Cartographic generalization is always applied on
  symbolized (cartographic) features

25
Cartographic generalization operators

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Enhancement
• Displacement
• Elimination
• Typification
• Amalgamation

26
Examples cartographic generalization operators

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

Displacement
Amalgamation
Elimination
Enhancement
Typification
27
Outlook

• Introduction motivation
• Classification framework
• Model generalization operators
• Cartographic generalization operators
• Outlook

• Verifying classification by interviews with NMAs
• Harmonizing model generalization and schema
  translation
• Research collaboration with Lassi Lehto _at_ FGI
• Extracting implementing a formal description of
  the operators

28
Conclusion

• Classification framework provides an extensible
  means to classify operators
• Initial classification is based on literature
  research
• Verification implementation still has to be
  done
• However the classification will always remain
  subjective (up to a certain limit)
• Classification is basis for formalization

29
Discussion

• Merge vs. Fusion
• Amalgamation vs. Aggregation
• Enhancement Enlargement

30
Thanks for your attention

• www.durpondergronden.nl
• foerster_at_itc.nl