Does Google Maps or OSM use any deep learning or machine learning model to extract walkways from satellite images? Examples of walkways are the green lines in the picture here. I wonder how they were added to the map in scale?

Most papers about pedestrian network extraction I found online are at least 7 years old. I wonder if there are any more recent work in walkway extraction?

UW campus showing walkways in green

  • 1
    Maybe in some cities, would be my guess. I know they use contractors to digitize & correct roads and pathways. I think some of the data comes from authoritative sources while some data is generated using data from street view cars and google maps users. Sep 22 '20 at 3:04

In this case the source is Open Street Map - you can get the history for the paths (walk ways).

Example: https://www.openstreetmap.org/way/4708373/history

you can find out that it was created 13 years ago and the 'Potlatch alpha' was the original source of editing.

Potlatch was a flash based editor for OpenStreetMap this has been superseded by Potlatch 2 architecture (Javascript) https://en.wikipedia.org/wiki/Potlatch_(software)

update the iD editor https://en.wikipedia.org/wiki/ID_(software) is the current software for editing in the browser


I would say most, if not all, of openstreetmap walkways are manually added from tracing satellite imagery or GPS traces. Iv'e never heard of anything automated regarding walkways in openstreetmap. Coastlines and landcover has seen extensive use of automation in some form, but that was mostly early in the project. Adding massive amounts of data today is a bit harder, since you need to not destroy anything.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.