Many cities don't have quality data on their walking infrastructure. Detailed and frequently updated walking network and asset data are usually entirely missing or collected in inconsistent formats. Alternative solutions like OpenStreetMap also suffer from many years of car-centric 🚗 focus with little information for pedestrians.
A research study led by Heidelberg University in 2017 flagged the incompleteness of the sidewalk data on OpenStreetMap in five major German cities including Berlin, Munich, Hamburg, Freiburg, and Heidelberg. The study revealed that only 22.5% and 17.6% of sidewalks have been mapped in OpenStreetMap with regard to the total number of sidewalks and total length of sidewalks, respectively.
Confirming the already-known limitations of OpenStreetMap in relation to walking infrastructure networks, here is a picture that's worth a thousand words from Sydney in 2023.
On the left, the Sydney CBD OpenStreetMap walking network data are shown that looks quite incomplete and patchy. On the right, our GeoAI-generated sidewalk routable network overlaid on OSM gives a much more complete picture of the state of the walking infrastructure. The sidewalk network data is generated using our proprietary street-level sidewalk view imagery and AI-powered mapping process that includes a visual semantic segmentation deep learning model and a monocular depth estimation model with transformer learning. The outcomes of the two models are then fed into a number of back-end algorithms to create a routable sidewalk network.
Want to learn more about footpath.ai and how we use AI to map sidewalks in a fraction time and cost compared to traditional surveying? or simply want to have your city's walking network mapped out and added to OSM, please get in touch. We'd love to hear from you.
