Background

Agencies such as the Red Cross, the World Bank, and Doctors Without Borders make use of OpenStreetMap in many of their operations. As a large VGI project, OpenStreetMap shows a variety of data quality in different areas. There could be errors due to many reasons: the area has not been digitized by anyone yet, no up-to-date high-resolution imagery was available, or the experience of the volunteer mapper. The agencies need to know in which areas they can rely on OpenStreetMap data and in where they potentially need to contribute or ask assistance to update the map data.

There are several methods to analyse OpenStreetMap data [1, 2, 3], see also Figure 1. Tools such as OSM analytics [4], for example, which allows us to do gap detection and compare the completeness with other datasets, such as the Global Human Settlement Layer [5] and WorldPop [6, 7]. Other tools make it possible to visualize aspects of OpenStreetMap’ mapping quality, such as version history and contributor activity [8, 9], see also Figure 2 and to know the age of satellite images [10].

Apart from analyzing the quality of exiting map data, the agencies also need to know whether there is new aerial imagery available to enable the mapping a more actual situation of buildings, infrastructure, etc. Again, there are tools to do such [11, 12], see also Figure 3.

Despite the usefulness of these tools, they are scattered and somehow distant from the people initiating the mapping campaigns. This project therefor aims at bringing some of the OpenStreetMap quality properties closer to them and the tools they are using, such as [13, 14].

Tasks

• Evaluate methods of OSM accuracy determination (based on feature edit history and contributor skills, date of image used, validation done, cumulative edits, etc.).
• Evaluate methods of OSM accuracy visualization.
• Design method to improve accuracy metadata during mapping -> organize mapping event with variety of image dates.
• Evaluate methods to search for new satellite imagery available.
• Present metadata to mappers and users as close as possible to their user interface (e.g., id editor or synchronized website)
• Describe an integrated approach with a combination of tools and guidelines.
• Organise small mapping event to test the approach.

This topic will be supervised together with dr. Caroline Gevaert (EOS department)