OpenStreetMap

Mapping roads and buildings in Singapore!

Posted by poornibadrinath on 24 November 2016 in English. Last updated on 25 November 2016.

In the last two weeks, as a part of improving the quality of base-map data of Singapore on OpenStreetMap, the Mapbox data team along with the community has completed adding missing streets and buildings in Singapore.

For this task, the team used a combination of Mapbox and Bing satellite imagery to improve the road network and building footprints data. We presently have managed to refresh more than ~1200 kilometers of roads. Additionally, we also have added close to ~40,280 buildings.

Take a look at the visualisation below which shows the buildings added during the last two weeks.

sg_builings

image

Issues with mapping in Singapore:

  • The main issue that the team encountered was the lack of good quality satellite imagery for both Bing and Mapbox. While Bing had better quality than Mapbox, it was not recent and had huge patches of cloud cover. So secondary sources like Strava were used for verification of major roads during adding missing roads.
  • Also due to the presence of huge number of high rise buildings, there often was distortion and the building footprint had to be very carefully aligned around the base and not the rooftop.

Singapore is a rapidly developing country and satellite imagery is often outdated. It would be great to have the community validate our edits and give any feedback regarding our process in our mapping ticket.

We thank the community for all the support. We look forward to more interactions with the Singapore community.

Thanks,

From Mapbox Data Team.

Discussion

Comment from BushmanK on 24 November 2016 at 20:53

Are you aware of this information? https://data.gov.sg/dataset/sla-cadastral-land-parcel License of it seems open enough for OSM: https://data.gov.sg/open-data-licence

It is not equal to building footprints or roadway centerlines, but you still can use it to check alignment and such things.

Comment from poornibadrinath on 25 November 2016 at 07:44

@BushmanK, thank you so much for sharing this. This is really helpful!

Comment from BushmanK on 25 November 2016 at 17:10

@poornibadrinath, it makes perfect sense to always start from searching for open data before dealing with any new area. Sometimes it could be not so easy due to language, but there should always be at least one member of a local community, right?

Comment from ashish nair on 7 June 2018 at 06:14

Hi @poornibadrinath any idea which imagery are you using for adding roads and buildings. i want to know which is the best imagery on OSM.

further what source are you using to find building names and addresses

Log in to leave a comment