Pretty much self-explanatory. I see the whole network of roads slightly misaligned with the satellite picture in the background (buildings also). I would start aligning them but it seems like such a generalized issue that I'm not sure if fine-grained editing is the correct solution.

asked 15 Jan '14, 15:00

gdiazc's gravatar image

gdiazc
31113
accept rate: 0%

edited 15 Jan '14, 15:35

aseerel4c26's gravatar image

aseerel4c26 ♦
32.1k16239551

1

Can you provide a link to an area where this is the case (or give a description)? There are a number of possible reasons, including that the aerial imagery is wrong, but knowing the area would help to determine the problem.

(15 Jan '14, 15:05) SomeoneElse ♦
1

Here, for example: http://www.openstreetmap.org/edit#map=19/51.75821/-1.25753. The curve of the grass and the road look slightly misaligned, and this continues for several blocks (I'm showing the University of Oxford).

(15 Jan '14, 15:10) gdiazc
2

Same area at z20 and it aligns http://www.openstreetmap.org/edit#map=20/51.75834/-1.25755 - one or the other set of images is slightly out.

(15 Jan '14, 15:26) EdLoach ♦

Oh yes. How can a problem like that be solved?

(15 Jan '14, 15:27) gdiazc

That seems to be "the" standard problem with aerial imagery (and multiple sources in general).

See e.g. bing-osm-and-my-waypoints-all-misaligned-who-do-i-trust or the others listed under the same tags (click on the grey words below the question and sort by vote count).

Preliminary advice: Do not move the existing OSM data to match the aerial imagery unless you are quite sure. We need to know the reality – find it out, somehow. Aerial imagery is not necessarily close enough (by today's standards) to the real position. The other way round: OSM data is not necessarily close enough (by today's standards) to the real position.

For discussing (evidence, hints, measurements, observations of) this specific problem the forum is likely the better place (at least for same questions regarding other locations).

permanent link

answered 15 Jan '14, 15:38

aseerel4c26's gravatar image

aseerel4c26 ♦
32.1k16239551
accept rate: 17%

edited 15 Jan '14, 15:41

1

If you zoom in, and then look to the west by St Giles, and then look at the underlying GPS traces

http://www.openstreetmap.org/edit?editor=potlatch2#map=20/51.75831/-1.26020

you can see that most of the GPS traces are to the east of the roads, suggesting an imagery offset in that direction at that zoom level. It's less clear north-south since there isn't a nearby way clear of tall buildings running east-west (St Giles is at least fairly wide).

If you want to contact people who've mapped Oxford I'd start on the talk-gb mailing list or the #osm-gb IRC channel (see @aseerel4c26's contact link).

(15 Jan '14, 16:06) SomeoneElse ♦
1

Another way to contact the local mappers is via the built-in messaging at www.OSM.org and finding out who mapped specific features.

(15 Jan '14, 23:24) aseerel4c26 ♦

To answer "How can a problem like that be solved?", once you've figured out (from a combination of various sources and local knowledge) how offset imagery is in a certain place, you can align the imagery in the editor. This varies editor by editor - for example, iD has a "fix alignment" option on the background menu.

permanent link

answered 15 Jan '14, 16:12

SomeoneElse's gravatar image

SomeoneElse ♦
32.2k63333752
accept rate: 15%

1

… It may also be (and should not be forgotten) that the imagery is not offset but the current OSM data is. That was the situation in my mapping area: the older aerial imagery before 2012 was offset (in addition varying offset due to differing ground elevation), and apparently many mappers used it without correcting the offset leading to offset OSM data.

Moving OSM data (if you/others are quite sure it is offset) is not easy as multiple objects need to be selected (best in JOSM), care must be taken of relative distances/angles, objects which are not offset must not be moved, …

(15 Jan '14, 23:22) aseerel4c26 ♦
2

… My "mappers used it without correcting the offset" sounds a bit reproachful. Not meant that way (by the way, I was one of these pre bing2012 mappers). (I think this applies to many locations:) in earlier times less gps traces were accumulated, the aerial imagery had less good resolution (more blurry when zoomed in), there was less OSM data, there was much work to do to get a basic street map complete to some extend. Investing much time in correcting relatively(!) unimportant offset issues may not have been rational.

(16 Jan '14, 03:00) aseerel4c26 ♦
Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Question tags:

×76
×36
×11

question asked: 15 Jan '14, 15:00

question was seen: 5,763 times

last updated: 17 Jan '14, 16:11

powered by OSQA