Corrdinates from GPSr compared with Google Earth coordinates

[+dfx]

from today's short canoe trip... oregon tracklog vs. google earth.

 

you can see the walkway that we walked down on the aerial image. the tracklog disagrees. it's off by about 70 feet. according to google, we went right through the building.

 

 

and as for the canoe trip itself, according to google we went through the trees!

 

Edited June 26, 2010 by dfx

[+thedeadpirate]

Google Earth is consistently 300' off here.

 

Recognize anything here? If your 300 ft is correct your cache is not anywhere in the photo.

 

 

There's an extra 0 in there.

 

Look at my Over The Flint cache and you will notice Google places the road and cache about 30' feet south of the actual location.

 

And the cache you are showing is actually located at the Northeast corner of the block right beside the sidewalk.

[robertlipe]

I understand rather well how Google Earth and related processes work and this thread is full of bad science. The process of taking photos and mapping them to an oblate spheroid is not a lossless one. Some corrections has to be done and often, some manual intervention is necessary. As the images are being constantly updated, sometimes it has to be done again.

 

If the imagery in your area is "consistently shifted N meters in direction X", report the issue at https://earth.google.com/support/bin/reques...ntact_type=data

[I!]

If the imagery in your area is "consistently shifted N meters in direction X", report the issue at https://earth.google.com/support/bin/reques...ntact_type=data

Thanks for the link. That should be a big help for those contributors suffering from misaligned imagery.

 

Do you happen to know what alignment accuracy Google would regard as acceptable in the U.S. and Europe?

 

Also, since you "understand rather well how Google Earth and related processes work" and "this thread is full of bad science", perhaps you could give us some further information about how the mapping process works in practice and what some of the common sources of errors are.

[robertlipe]

Accepted consumer-grade GPS accuracy/precision is around 3 meters. I wouldn't bother reporting a problem because the imagery said you're in parking spot 46 while your GPS said you're in parking spot 47.

 

Satellite/aerial imagery really isn't meant to tell you which evergreen tree a bison tube is hanging in. That's just not a reasonable expectation.

 

Images taken at high altitude are very focused at the center and are "warped" at the edges. But a consumer rarely knows where the center of said imagery is. Nerdy math can compensate somewhat for that, but it works sometimes better than others. (Curse this planet for being neither flat nor spherical!)

 

As the "best" imagery for any area depends on a lot of things (perhaps we were closer to the area you care about yesterday, but had 30% cloud cover - that's probably less "good" than the imagery today that's further away from the center of the lens, but with 3% cloud cover) there are a lot of factors involved in deciding which imagery is the "best".

 

If a number of people independently submit a GPX file with GPS tracks driving the same road that's 1000 feet NNW from where the imagery says that road it is, it's not too hard to convince the imagery group that those bits are broken and should be fixed.

 

Disclosure: I've been on the Earth team at Google for over three years, albeit not in the imagery team.