1. Generalisation process and generalisation tools in Maanmittauslaitos
Ari Öysti
Small scale map production

2. Automated generalisation process
Operator starts the batch process
Data flows from tool to tool in automated batch process.
Input data
Generalizedoutput data
Tool 1
Tool 2
Tool 3
Is this really possible in complicated generalisation process ?
Yes it is possible: e.g. fully automated generalisation process in Holland Kadaster 1:1000 -> 1:50 000

3. Interactive generalisation process
Data flows from tool to tool controlled by operator. Operator checks the validity of the input data and output data in every phase of the process. Tools can’t do everything. You need also manual interactive work during the process e.g checkings and corrections.
We use this kind of process in our generalisation process.
Tool 2
Generalizedoutput data
Input Data
Tool 1
Tool 3
Operator starts tool
Operator starts tool
Operator starts tool
Operator starts tool

4. Tools are not enough, you need to desing the whole generalisation process = generalisation process model
What generalize operations = tools are used for each map theme. First thing is to define the generalized output product.
In what order map themes are generalized.
What controll parameters are used for each operations = tools for each map theme.
Does the different map themes have topological relationships which should be maintained during the process.
Is it necessary to spend too much time and money for producing over-quality in small details if enough-good product is ok for the users. Small scale maps are usually used as zoom-in and zoom- out maps.

5. Dataset scales Cartographic Names Topographic layers
Topographic DataBase
1:10 000
Name DataBase
1:25 000, 1:50 000
JAKOmtj
Smallworld
Map 1:25 000
Map 1:50 000
Transfer from Smallworld to ArcGis in shape format
Generalisation
Manual generalisation 1: 1:
2008
International datasets
Generalisation
Manual generalisation
ERM, EBM 1: 1:
2008
Generalisation
Manual generalisation
1: , 1:
2011
EGM 1:
2011
Manual generalisation
Generalisation
1: ,
1: , 1: 1:
2011
2011
ArcGis/Geodatabase
JAKOmtj/Smallworld

6. Map theme layers are generalized separately in own generalize processes. Topological relationships between themes are maintained.
Hydrography theme
Rivers
Land use and vegetation theme
Lakes
Agricultural areas
Sea
Wetlands
Transportation theme
Rocks
Roads
Administration theme
Railways
Boundaries
Airports
Areas
Settlements theme
Elevation theme
Buildings
Contour lines
Built-up-areas
Cartographic names theme

7. Generalize processing is done with 5 PC’s by 5 persons in Pasila and 1 person controlling the workflow
Data is divided into working areas and working areas are processed at workstations. After processing these working areas are merged and edgematched to seamless dataset.
1: (vector format 80 Gb)
1: (vector format 3 Gb)
1: (vector format 600 Mb)
1: (vector format 80 Mb)
1: (vector format 5 Mb)
e.g. Transportation theme from 1: > 1: takes about 6 months and 1: > 1: about 3 months
Processing 1: and 1: datasets is done whole country as one working area

8. Main use for generalized maps are zoom-in and zoom-out orientation rastermaps in our own mapservices. This is probably the main use how the customers use these datasets. We don’t make printed maps from these generalized datasets

9. Generalized datasets and our own rastermap products in our mapservices
Topographic map 1:
Background map 1:40 000

10. Generalized datasets and our own rastermap products in our mapservices
Topographic map 1:
Background map 1:
Background map 1:80 000

11. Generalized datasets and our own rastermap products in our mapservices
Topographic map 1:
Topographic map 1:
Background map 1:
Background map 1:

12. Generalized datasets and our own rastermap products in our mapservices
Topographic map 1:
Topographic map 1:
Topographic map 1:

13. Generalized datasets and our own rastermap products in our mapservices
Background map 1:
Background map 1:
Background map 1:

14. Own generalisation tools = PIEKKA application PIEKKA = PIEnimittaKaavaiset KArtat
ESRI Finland’s GIS-Award of the Year 2011

15. PIEKKA application
Developing of the application started 2003 (3 developers, 2 test people)
New generalisation production system in use since 2006.
Basic idea for generalisation process: Avoid interactive generalisation work and use automated tools when it is possible
Tools in standard GIS-applications are too simple for high quality generalisation
Developing own generalisation tools was a big challenge.
Based on ESRI ArcGis application and ArcObjects componets
ESRI’s ArcObjects components provides a wide component library for developing own generalisation tools
Own tools are integrated into ESRI’s standard ArcMap application using ArcGis addinn setup file
Database format is ESRI ArcGis FileGeodatabase format
Topological relationships between features are maintained in FileGeodatabase with ArcGis topology tools

16. PIEKKA application
PIEKKA generalisation tools are integrated to ArcMap application

17. Generalisation parameters
PIEKKA application
Generalisation parameters
All parameters for controlling generalisation process are stored into separate control database tables.
Application reads the generalisation parameters from the tables of the control database.
All data is generalized in the same way and users can’t use wrong parameter values by mistake.
You don’t have to rewrite the tool’s code if you want to change the parameters.
Same tools can be used for generalisation of different scales.
Typical cartographic visualisation and symbolisation rules in each map scale defines the values of generalisation parameters.

18. Generalisation parameters
PIEKKA application
Generalisation parameters
Same tools can be used for different levels of generalisation by switching the generalisation parameter database.

19. Generalisation parameters
PIEKKA application
Generalisation parameters
Same tools can be used for different levels of generalisation by switching the generalisation parameter database.

20. Generalisation parameters
PIEKKA application
Generalisation parameters
For special cases user can set generalisation parameters by using tool’s user dialog forms

21. Generalisation methods
PIEKKA application
Generalisation methods
Generalisation operations available with customized tools:
Simplification
Collapsing
Enhancement
Selection
Elimination
Displacement
Aggregation / Merging
Generalisation application has 35 tools to run generalisation operations listed above.
Usually tool for running one operation consists of several base methods

22. Base methods for polygon aggregation operation
Piekka application
Base methods for polygon aggregation operation
Narrow sections
Create merging area between polygons
Create centerline into merging area
Clip the merging area by using the centerline and join together the clipped parts into neighbor polygons

23. PIEKKA application Examples Selection by attributes of features
by geometry (length or area size)
Feature type collapse
polygon  line
polygon  point
line  point

24. PIEKKA application Examples Line and Area simplification
by Douglas-Peucker algorithm

25. PIEKKA application
Examples Type collapse: Polygon -> Line Narrow polygon river -> line river

26. PIEKKA application Examples
Displacement of buildings lying near the road line
Typification of dense building clusters (= maintain the general pattern of building group shown in their approximate locations)

27. PIEKKA application
Examples Buildings are not moved over water areas or other side of the road or railway or river.

28. Line generalisation PIEKKA application Examples Line fusion Line merge
Line sections under minimum length are joined together with the longest neighbor.
Line merge
Separate lines near each other are replaced by creating a centerline between original lines and transfering attributes to new line
Removing under minimum length branch lines
If there are several under minimum length branch lines at the same node the longest one is kept and the other branch lines are removed

29. PIEKKA application Examples
Merging and removing rules for small area features with different land use classes.

30. PIEKKA application
Examples

31. PIEKKA application Examples
Rules for merging and simplifying area features

32. PIEKKA application
One of the most challenging task for automated generalisation = complicated road junction
1:10 000 1:
We do this manually!

33. Thank You