People should be wary of the imagery offset especially using high resolution products. In OpenStreetMap, Mapbox often provides high resolution images in densely populated areas within the US. However, just because Mapbox provides high resolution imagery does not mean the offset is not a thing. In fact, Mapbox has created a lot trouble as it pretends to be accurate but actually not.
Keep in mind that whenever your local municipality has its own orthoimagery published, always use it to draw features (of course it has to be compatible with OpenStreetMap, if it is missing in iD, maybe you can bring it in). The imagery could be dated, but it is accurate. Your local municipality will spend more time to process with the data. These “commercial” providers, like Mapbox or ESRI or Bing, do not have a lot of time to carefully examine everything in a specific area especially when it is not that popular.
I’m going to post an example here. The orange area is OSM data fetched by QuickOSM in QGIS, drawn by a random person who has not noticed the offset for years.
Mapbox imagery (inaccurate but free to use in OSM):
MD State-wide imagery (accurate and free to use in OSM):
Engineering imagery (accurate but cannot be used for OSM):
MD State-wide imagery is a little bit different from engineering imagery because of different angle, they are de facto the same.
The above example is from Maryland of the USA. In Maryland, please find
MD Latest 6 Inch Aerial Imagery as it is always a standard!
Comment from andy mackey on 10 October 2021 at 09:44
Using all available GPX traces, in an editor, will allow us to check and adjust imagery. Several traces will give a fairly accurate result. Even if an area is on OSM it may be useful to upload GPX data because the area has been mapped by a mapper who has been fooled by misaligned imagery, you see it and you believe your own eyes :-) so upload those traces. If your device seemingly off sets your position a similar amount on the map several times, that bit of map may have been created from off line imagery. Happy mapping