• 

  User can select an issue...

• 

  ...and take a picture of it.

OSM Runner

We are inspired to improve the open street map data with a new approach in user interaction and image classification using machine learning. Users can collect points by taking pictures at different spots. There are plenty of them spread across Munich.

How it works

We build an Android app showing all the open issues on a OpenStreetMap map. Users can select a issue which leads to a camera view coming up. Afterwards the picture is being sent to the server via Rest request. The server is also able to provide all the issues in a specific area to improve performance with great amounts of issue points. When receiving a picture the server is analyzing it using the trained neural network to generate tags.

Gamification

The app currently has a simple leveling feature. After collecting enough issues the level improves. To improve comparability the level should be synchronized via server an a global position should be shown. As a longterm feature the user should be able to specify a kilometre count and the app tries to create a running track covering as many issues as possible.

Built With

• android
• java
• postgresql
• python