Eastern DRC & Western Uganda Maxar-Ecopia Buildings Import

From OpenStreetMap Wiki
Jump to navigation Jump to search
Eastern DRC & Western Uganda Maxar-Ecopia Buildings Import - Other languages
Main page Workflow

The Eastern DRC & Western Uganda Maxar-Ecopia Buildings Import is an import of building footprints covering the north-eastern part of the Democratic Republic of the Congo (Ituri and North-Kivu provinces) and the western part of Uganda. The import is currently going through the Import Guidelines steps, and it should be soon an ongoing crowdsourced task.

Goals

The Humanitarian OSM Team and its community has the goal of importing these building footprints into OSM in a crowdsourced way, merging them with the existing data, improving their geometries when needed, and dismissing those that aren't correct - using aerial imagery for validation.

More detail on the context of this project can be found on this page.

Schedule

  1. Preparation, discussion - started May 11th 2020.
    1. Discussion with DRC & Uganda OSM community and HOT list: started June 23rd 2020.
    2. Discussion in the Imports list: from July 20th to August 3rd 2020.
  2. This data import won't be just a blind import, but a crowdsourced import, where users are expected to check the validity of the data against aerial imagery, and correct the geometry of each building according to those images. Any user is welcome, but a good experience with OSM mapping (buildings, POIs, overlaps, offsetting imagery) is required.
  3. Bearing in mind the quantity of data, this import will be split in several files and be done one file after the other. So we expect it could take several months.

Import data

Data description

The original data is in shapefile format, and consists of around 6.6 millions building footprints across the north-eastern part of DRC (Nord-Kivu and Ituri provinces) and the western part of Uganda. MAXAR has given permission to import the data into OSM.

Data has been generated by automated feature extraction based on MAXAR imagery performed by Ecopia AI and funded by the Humanitarian Grand Challenge. The main feature extraction and data validation was mainly performed using training sets from the area etc.

All building footprints could be imported into OSM, except when existing buildings were traced or created from a more recent source. In the cases where there are existing buildings footprints traced from older imagery, the MAXAR-Ecopia building footprints will replace the older ones, keeping the history as much as possible, and hence improving the OSM database.

Background

Data source site:
Data license: MAXAR and Ecopia.AI have given permission to import the data into OSM, by internal documentation.
OSM attribution: https://wiki.openstreetmap.org/wiki/Contributors#HOT_.28Humanitarian_OpenStreetMap_Team.29_crisis_mapping
ODbL Compliance verified: yes

OSM Data Files

For concrete examples of Maxar-Ecopia building footprints data, you can check these files 1 and 2 used in test projects set up in the Tasking Manager.

Import type

The best approach to import this data is manually by the community, with data split in a number of files, that will correspond to an equal number of projects in the HOT Tasking Manager. Advantages:

  1. Any volunteers may join the import effort, at any time (although only skilled mappers should join).
  2. We can check, at any time, the mapping progress of the import.
  3. We can easily validate each task too, and check the validation progress.
  4. Easy to set up.

We will provide a link for each Tasking Manager project in a table of the import workflow wiki, with info about the total size (number of buildings) and expected time of work for each one. We will open the different projects progressively, when the previous jobs are being finished.

Data preparation

Data reduction & simplification

No simplification of the original data is expected.

Tagging plans

Some of the tags of the original dataset aren't relevant, so we ignore them. Here we list all the original tags with their corresponding translation into the OSM tagging schema, plus additional tags for all segments:

Maxar/Ecopia key Maxar/Ecopia values meaning OSM tag Comments
ImgDate Date of the imagery used to extract the building footprint source:geometry:date=*
DataSource=Ecopia Tech Maxar Source of the data source:geometry=Maxar_private;Ecopia.AI Renamed
CAT_ID We ignore it
PolyDate Date of the building footprint extraction We ignore it
Building tag for new objects building=yes Original tagging is maintained for existing buildings

Data transformation

  1. The original file (in shapefile format) was opened with JOSM + Open Data plugin, and saved in osm format.
  2. We deleted all non-relevant tags and translated the info of the rest to produce the proposed tags.
  3. The resulting file will be divided in pieces, one for each Tasking Manager project, again with the JOSM editor.
  4. Any errors detected by JOSM Validator (empty relations, duplicated nodes or ways...) are fixed in the files before publishing Tasking Manager projects.

Data transformation results

You can download the data files for the test projects set up in the Tasking Manager: the first and the second.

Changeset tags

We will use the following changeset tags:

Tag automatically added by the Tasking Manager:

where NUMBEROFPROJECT is the Tasking Manager project number.

Tags to be updated or added manually:

Example:

Data merge workflow

Team approach

This import (data integration) will be done through the HOT Tasking Manager, so the number of people importing the data is unknown. We don't expect a big number of users, as we require them to be experienced mappers. Among the skills required:

  • Good experience with JOSM.
  • Experience with previous imports through the Tasking Manager is not necessary, but a plus.
  • S/he knows how to use JOSM filters.
  • Skilled working with buildings. It's needed to know well how to orthogonalize buildings (Q), round buildings (O), combine ways (C) and ungluing (G).
  • S/he knows how to use the Replace Geometry tool of the UtilsPlugin 2 JOSM plugin (Ctrl+Shift+G), and why it is so interesting.
  • S/he knows how to use the ToDo JOSM plugin.
  • S/he knows how to create and deal with building relations.
  • S/he knows how to deal with conflicts.

References

Following the imports guidelines, this import has been discussed first in the local DRC OSM mailing list, local Uganda OSM mailing list, and to get more feedback it was also shared with the HOT mailing list.

Workflow

We will use the JOSM editor for this import.

Step by step instructions

There are several ways to reach the same results. We have tried to find the simplest one, the less error-prone and the one that assures a higher level of consistency across different import volunteers. The proposed workflow is as follows:

  1. Choose one task of one of the TM import projects. User will be presented with 2 sets of data (the OSM data and the new Maxar/Ecopia buildings to be imported) for the task square in two different layers, as well as the imagery layer which was used to extract the buildings.
  2. Correct any significant offset between the existing OSM data, the project imagery and the Maxar-Ecopia data.
  3. Depending on the absence or presence of existing buildings in OSM in the selected task square, the contributor will either merge the 2 data layers and add all buildings to the todo list, or configure the conflation plugin.
  4. Review each building individually going through the todo or conflation lists and decide on the geometry replacement, addition, or correction.
  5. Upload the data into OSM with your import dedicated account.
  6. (Optional) If a certain amount of missing buildings have been spotted during the review of the area, these can be manually traced and added with your regular OSM account.

Changeset policy

Changesets will be small in size, so no issues expected in this respect.

Revert plan

If something goes wrong (quite unlikely based on experience), JOSM reverter will be used.

Conflation

Conflation is explained in the import workflow wiki, in the sections 3.2.2 and 3.3.

The import workflow will basically keep the tagging and the geometries of the existing OSM buildings, unless the Maxar/Ecopia equivalent is of better quality or simply more correct. All existing tags will be preserved.

If an existing building in OSM is no longer visible on the most recent imagery (theoretically the one provided by Maxar unless the existing building was traced on customized imagery), it will be removed unless it contains any valuable tags but a fixme tag could be added to it.

We will also make sure buildings and highways don’t cross each other.

Acknowledgement of other organized editing and imports

Since several mapping teams are currently active in the region, we acknowledge being aware and coordinating with other teams including:

  • UN Mappers, mapping landuse in the surroundings of Beni and Butembo (Nord-Kivu, DRC)
  • MSF regarding Lopa & Lingo buildings import (Ituri, DRC)
  • Beni and Butembo OSM contributors (Nord-Kivu, DRC)

QA

Validation of the import will be done by a second user in the Tasking Manager.

Retrieved from "https://wiki.openstreetmap.org/w/index.php?title=Eastern_DRC_%26_Western_Uganda_Maxar-Ecopia_Buildings_Import&oldid=2124584"