Measurement of Safety in public spaces through Street Viev Imagery

Purpose and goal

This project aims to develop and test a platform for measurements in an urban environment - of e.g. places´ safety, accessibility and/or traffic safety with the help of image analysis ("Computer vision") on Street View Imagery. The images can be user-generated, come from Google Street View or other large-scale image platforms and constitute a huge and in Sweden rather underutilized data source for measurements in cities. The project intends to test the technology as a proof-of-concept.

Expected results and effects

** Denna text är maskinöversatt ** The expected result of the project is that RISE, together with the owner of the needs, gets to test the technology, and can assess the potential for scale-up. We expect that the utility and potential of this technology can be very great, but it needs to be shown in small format first. If a measurement about a specific place quality for a municipality has been made and a neural network trained for this, the same network can be used by another municipality for similar measurements.

Planned approach and implementation

The project will - gather images through browser emulation) over the urban area of Tomelilla (which is a pilot municipality) from image platforms - using computer vision techniques extract information about specific place qualities associated with perceived safety - based on the above technique, answer specific questions about places in an area: e.g. "Are there one or more conditions on the site that are associated with the feeling of insecurity?" - with the technique described above, be able to map relevant site qualities over a more electric. smaller larger area in a city