Automatic pedestrian crossing mapping for our road safety

The Road Safety Data Taskforce was established in early 2020 and consists of a network of governments and knowledge institutions. The Taskforce is working on a risk-based approach to road safety. One of its wishes is to gain insight into the exact location of pedestrian crossings, or pedestrian overpasses (VOP). NEO has automatically mapped crossings based on aerial photos (open data). This allows for better analysis and helps road authorities improve road safety.

Road safety in the Netherlands is still a hot topic. By 2050, the aim is to have zero traffic fatalities. By comparison, there were 610 traffic fatalities in 2020, of which 41 were pedestrians. Most serious road accidents involving pedestrians take place within built-up areas with a car as the other party [1]. The exact location is often not recorded, while understanding this location is important to take appropriate action. To improve this registration and enable data analysis, it is crucial that road characteristics are properly recorded. That is why the Road Safety Data Task Force is collecting more and more data on these road characteristics.

One of these road features is the presence of pedestrian crossings, known to most people as zebra crossings. The locations of pedestrian crossings are not always recorded. This is where NEO can help. NEO has applied artificial intelligence to automatically map pedestrian crossings based on high-resolution aerial photographs. We did this through a deep learning model that uses shape, texture and colour, among other things, to recognise objects. We succeeded in training this deep learning model to detect zebra crossings even in difficult situations where, for example, trees or cars partially obscure the view of the zebra crossing. The location information can then be linked to the National Roads Database (NWB) and the Road Characteristics Database (WKD), allowing for better risk-based road safety policies.

For more information, please contact Steven Braakman (steven.braakman@neo.nl) or Emile Nieuwstraten (emile.nieuwstraten@neo.nl).

Source:
[1] SWOV (2020). Pedestrians. SWOV fact sheet, July 2020. SWOV, The Hague, the Netherlands.

We care about our planet, which is why we keep a close eye...

NEO captures objects from above and monitors changes smartly and accurately. In doing so, we create data and insights that directly contribute to your challenge.

Curious what that will bring you?

Also interesting...

Updates data services in one place from now on

Green-Blue Veining: what it is and why it matters

NEO
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.