Do surveyed points matter?

General talk
1

From time to time, I do have access to surveyed data which is about 30mm accuracy. Few questions..

Do imagery providers use them as ground control points whenever available? They seem to have different projection ratio even when centered at a common point.

More importantly,

Do we shift ways and nodes in OSM to correspond with the new data? what radius should be involved? will this become an endless endeavor since not all editors have access to image alignment database? My personal opinion is anything within 3 meters (the accuracy of our handphones) should be left alone, although my OCD side seems to disagree much.

regards.

1 Like
2

it depends:

if you edit with smartphone GPS/unaligned aerial: yes, you should not make adjustments subtler than available accuracy

if you have access to better data and feel like doing such adjustments - it is fine (I would leave decision whether do that or other OSM mapping or other OSM activity or something else altogether to a given person).

Though I probably would not encourage it

3

Normally yes, that’s their role!

4

Also worth considering that (from what I can tell) the vast majority of contributors don’t have access to or know about this type of data, so there is a good chance that if you put lots of work into high-accuracy adjustments, it will be unintentionally messed up over the years. Most mappers just rely on aerial imagery for alignment.

2 Likes
5

If you’re placing a pin based on this kind of data you should probably put source and note tags to indicate how it was placed.

Doesn’t always stop someone moving it, but at least there’s chance it’ll get left alone.

3 Likes
6

Also if you put a source and/or note, if it does get koved, we can check history of the object to try to determine what is the canonical position

7

that is exactly what I think will happen. between newbies, hot osm, etc etc, everybody will realign ways to whatever imagery they’re using. endless realignments. I think I’ll just put the markers in to see how far off everything is in the imagery, and only fix the really bad ones. esri and mapbox accuracy in my country is getting pretty good nowadays.

8

yes they do matter.
But
if there is a bin left to a street, and the bin’s exact lat/lon would mean to draw the bin on the right side of the street ('case the street is not so exactly positioned)
I’d rather place the bin on the street’s left side than at its exact location.

2 Likes
9

Exactly, never move surveyed points (except if you really know what you’re doing). Align the rest if you can. And add stuff relative to old stuff if you can’t.

1 Like
10

just finished a survey 2 days ago. I am rather impressed that Esri imagery is spot on at two points of building edges ~10 meters apart, but 200+ meters away the imagery already shifted by almost 2.5 meters. well within ‘tolerance’ i’d say, good enough.

thanks for all the replies.