I’m wondering what is the best practice when adding a path based on 3D imagery without seing it in real life.
Context
Sometimes I see some paths on a 3D imagery that are very likely missing on the map. This is most of the times in remote location or in paths which are not popular, like dead end paths.
Question
Would it be a good approach to proceed as follows:
Add the path, specify the source in the changeset (3D imagery). Whenever I’m not 100% sure, adding a “fixme” attribute encouraging to check the path.
Motivation
The reasoning is that it’s better to make a contribution that is 90% accurate rather than not making any contribution at all. In the worst case if the path is abandonned, then it can be marked as such once someone checks it.
EDIT: I’m using the available satellite photos, Strava heatmap and traces to gather as much evidence as possible. Usually the paths are in regions with low popularity.
If you know that the path exists near that location it is ok to use elevation data to align it more accurately. I wouldn’t use elevation data as the only source for the existence of a path. There are a lot of abandoned logging roads, etc. that show up in the elevation data. Many of these are very over grown.
I understand. If this is the general consensus of the community then I will refrain to map such paths when I haven’t been able to see them on the terrain.
Just wondering: adding the path with the fixme note would help future mappers. They can then just check it in the terrain and if the path is overgrown, then it could be tagged as abandoned or marked as vegetation as obstacle. Future armchair-mapper will then see the abandoned path and will not remap it.
Although that sounds like a good idea on the face of it, most data users/consumers do not look at the fixme tag (only other mappers and QC tools tend to look at it.), and thus a hiking app (for example) may route someone down one of these paths, and if that is a critical part of someone’s journey (necessary to complete a loop without back tracking a great distance), and the trail/track is impassible due to overgrowth, they may find themselves in a bad situation.
Yes, if you know that the path is ‘abandoned’ this is a good idea in my opinion.
There was a mapper around my area adding forest tracks without local knowledge by copying them from such digital surface models. In lack of detailed knowledge they just used “highway=track” for it, no further attributes. On the ground most of these tracks proved to be
out of use since long time and nearly unvisible due to vegetation, fallen trees and other kind of deadwood covering them. I do not think adding such tracks without any local knowledge by just copying them from 3D imagery does add much value to our data, and a fixme-tag would not improve that.
If such track is visible on aerial imagery or has some other support it is a different story but 3D imagery as single source is not sufficient imho.
Hi. I use 3D imagery (lidar) ALWAYS in combination with Strava heat map traces or in combination with orthophoto map. If at least two sources implies there could be a track then I map it.
I’d suggest that if you’re going to do that then please do make it abundantly clear on the changeset that “survey” wasn’t one of the sources used.
Personally I wouldn’t map anything that a router might send someone over if I didn’t know that it was actually possible to walk that path. As an example, a private dead-end into a field that I’d only seen one end of I might add from imagery, since I know that no-one (including me) is allowed to use it, but something that a router might suggest I wouldn’t until I was sure that it was actually a path all the way through and there were no nasty surprises.
Obviously this is a little different when you’re (for example) mapping after a natural disaster and what paths might be available is useful, but in an area where survey is possible, it’s different.
