[Imports] Import of Flemish Government data (building footprints and addresses)

Pieter Vander Vennet pietervdvn at gmail.com
Sun Oct 28 17:46:53 UTC 2018


Hello Import Mailing list,

Here in Belgium, we are planning to integrate building footprints into OSM.
The footprints are provided by the Flemish government under an open data
license. One of our community members built a tool to facilitate this
process: the tool automates merging the new data of different sources with
already existing data in OSM, after which a volunteer can load it in JOSM
to confirm and upload. The integration itself will happen in small,
supervised batches.
Our entire community is enthusiastically waiting for this integration to
get started.

With this email, we would like to inform the global community, seek
feedback and adhere to the import guidelines. More details on this i
ntegration are given below.

*# Data sources and licenses*

The data that we are planning to integrate is merged from multiple sources
published by the government (namely GRB, CRAB, 3D-gebouwenbestand). These
are all published under an open and compatible license[1], which explicitly
allows all reuse, including commercial. There is a need for attribution to
the AIV (Agentschap Informatie Vlaanderen - Agency of Information of F
landers). In earlier contact with them, they stated that a source tag and
their mention on the OSM Wiki[2] is sufficient.
 The license was already verified as being compatible with the ODBL earlier for
the import of stand-alone addresses (CRAB ) from the same source.

The government offers multiple data sets:

- GRB (Grootschalig Referentie Bestand - Large-Scale Reference File):
Contains geometry on buildings; as measured by geographers. Very high
quality in most places.
- CRAB (Centraal Referentie-Addressen Bestand - central addressess
reference file): Contains all addresses in Flanders.
- 3D building data: a file containing heights, etc. of all the buildings.

Note that all these files only are about Flanders – around half the surface
area of Belgium.

Note that these datasets are in use already for years as background layer
in both JOSM and iD. Most of the buildings and roads are thus already
aligned to them.

[1]
https://www.agiv.be/~/media/agiv/producten/grb/documenten/grb%20open%20data%20licentie.pdf
[2] https://wiki.openstreetmap.org/wiki/Contributors#Belgium

*# Community support*

There is a widespread support for this integration, especially amongst the
active community members. Not only will OSM contain _all_ the addresses
from Flanders once the integration is completed, it will also have accurate
geometries for each building. As a bonus, other actors (such local
municipalities) will also be more likely to trust and start using OSM
internally, as the governmental data will be already there.

A lot of active members support this project (such as Joost Schouppe, Xivk,
Polyglot, Glenn Plas, JBelien, Midgard, Tim Couwelier, Seppe, Escada ...)
Although technicals have been discussed thorougly, nobody is opposed to the
integration itself (especially because of the benefits).

*# This integration is necessary*

As the data is freely available, this data has been leaking into OSM for
years now, by various actors.
Some have manually traced hundreds of buildings, others have obtained the
shapefiles and added them to OSM – sometimes government officials themselves
!

By having the integration process in place, we can make sure that the data
is properly integrated, properly attributed and backreferenced for future
updates.

*# Process*

Glenn Plas has been building an awesome tool that combines the best of all
the datasets. If a building is already present in OSM, then the tags of OSM
are kept and the GRB geometry is taken.

If no building is already there, the tool tries to estimate the best
fitting tags depending on the size of the building, the landuse, ... These
speed up mapping tremendously but are meant only as suggestions to the
mappers.

Furthermore, the tool is not intended to 'simply add everyting at once',
but rather to map on a street-by-street basis: a mapper loads the data in
the tool, crosschecks with OSM and confirms the integration. The data
loaded in one such cycle are around a few streets. This way, manual
verification is done while the integration is going, preventing accidents
from happening.

The tool would be used by a select group of community members, who would
apply these integrations in locations that they know. We may even organize
one or more mapathons for the integration.

*# Documentation*

The process has been [documented extensively on the wiki](
https://wiki.openstreetmap.org/wiki/GRBimport).
The tool has been tried and tested already; Glen is working on even more
improvements.

*# Wrapping up*

All things considered, this integration is going improve OSM in Flanders
tremendously. Two years (!) of thought, discussion and development have
already gone into the tool, so we feel that the time has come to actually test
the tool on some scale.

We are looking forward to your feedback.


With kind regards,
Pieter Vander Vennet
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20181028/8f27848a/attachment.html>


More information about the Imports mailing list