Mapillary is the new Sweden based crowd courced Street View alternative that has taken the Open Street Map world by storm. One of the main feature according to their legal page is that all the photos are available under a Creative Commons Attribution-ShareAlike 4.0 International License with special permission to derive both photos and GPS data for use with Open Street Map. This post contains a brief writeup of a workflow I use to input information into JOSM.
TL;DR: line up the editor and mapillary using aerial imagery for best result.
First in Mapillary click on the layer icon on the right hand side of the screen under the + – buttons and enable MapBoxSatellite
This will change the Mapillary view to a Mapbox Satellite view which is significantly more useful than the OSM or MapBox view when dealing with a sparsely mapped area such as the one being edited. You should now get something that looks like this:
Look for a visually distinct landmark that you can reference in JOSM. You can do this by looking near the green pin for a unique road pattern, rooftop or some other feature that stands out. In this case a cluster of houses was chosen.
In JOSM, enable MapBox Satellite Imagery and remote control
and pan and zoom until you get to the same area (a short rant about this at the end of the post) and keep a look out for the reference points that you’ve chosen. UPDATE 2014-08-20: under the image there is a new icon labelled “i” if you click on it, right at the bottom, there’s a link that will open up the area in JOSM remote control.
Based on the reference point I could establish where the road to my left would be and I proceed to add it as a way. Finally, in the Mapillary interface, zoom in to the street sign to establish the name and number of the street.
And that’s another area successfully mapped in OSM!
One caveat is that if the area has low res aerial imagery then the other technique is to use the Mapillary2JOSM tool. I personally don’t like it because it loads up the imagery from the server with every click, a process which is not as smooth as navigating using the Mapillary interface itself.
Another caveat is if you’re referred to a Mapillary imagery, there is at present (2014-08-12) no way of knowing the coordinates of what you’re looking at and you will need to zoom in and out until you get it right. The issue has been opened at Mapillary’s bug tracker. Update 2014-08-20: There’s now a link available by clicking on the “i” under the image for JOSM and iD users to go directly to the area on the map.
Copyright © 2014. All Rights Reserved.
One thought on “Mapillary workflow for JOSM and Open Street Map”
Thanks for spelling out your workflow.
“JOSM and iD users to go directly to the area on the map” is a problem for me.
I don’t want to go to the “area” on the map. I want to go to a new node or cross-hairs or some other explicit location on the map.
I’m working in the woods and there are no nearby reference points. Just trees.