GPS accuracy and Google Earth

[+OpenTrackRacer]

I was just taking some readings today in preparation for creating a cache. I made five measurements with my 60CSx of about four minutes each and got four different results. This is not unexpected. I also noticed that two out of four times I looked with the GPS for the location, it pointed me off in the wrong area. Looking at the satellite page I could see that the receiver was showing different coordinates than the other times.

 

I took my numbers and put them into Google Earth. The location I'm using is easily seen on the satellite views. Google Earth shows a slightly different position than just about all of my GPS readings.

 

I'm wondering if anyone would care to comment on the accuracy of Google Earth's coordinate system? I could take the "X marks the spot" coordinates from Google Earth in this case if I wanted to. Would that be better or worse than taking what the GPS has provided?

 

Thanks!

[SiliconFiend]

I'm wondering if anyone would care to comment on the accuracy of Google Earth's coordinate system? I could take the "X marks the spot" coordinates from Google Earth in this case if I wanted to. Would that be better or worse than taking what the GPS has provided?

 

Thanks!

Probably worse. It's not that the coordinate system used by Google Earth is bad, but the imagery may not line up perfectly. For geocaching, all your coordinate systems/datums should be set to WGS84. That's the default all around for your GPS and GE.

[Grasscatcher]

Google earth is mostly designed/ slanted toward the "Oh Looky, there's my house" crowd. Don't bet on it's coordinate accuracy.

 

Let your GPS "settle in" and get a good lock and "average" before saving a point. Then save a second point at the same location. Compare ...and if they agree then you should be OK. If not, save a third one and it will probably agree with the 2nd.

This is assuming of course that you have a good clear view of the sky.

 

Generally, the most inaccurate coordinates are the first ones you save after turning your unit on.

Edited April 9, 2008 by Grasscatcher

[+OpenTrackRacer]

Yeah, the alignment with the imagry with the coordinate system was my main concern.

 

I was a little surprised and the differences I saw. It seemed to be off more than any time I ever looked for someone else's cache. A couple of times it was spot on and other times it was way off.

 

I'll have to mess around a little more.

[earthshaker]

Just for fun today on the way to work I marked a recognizable spot near the highway, with my Merigreen, I just checked the mark with GE, the location I marked was within 5-10 feet from where I actually made the mark, How did I come up with the estimate of accuracy? I stopped next to a Historical marker and took a picture of it after marking the location from within my vehicle, I got out of the Jeep walked around the front, went to the passenger door leaned my backside on the passenger door and took the picture. Since my jeep is about 5 feet wide and my GPSr bracket is mounted on the center console on the passenger side, which by my estimation was about 7-8 feet from the marker. The accuracy was not too bad I would say, of course your mileage will vary.

[+OpenTrackRacer]

That sounds similar to what I found. Google Earth seems to be five or so feet out of alignment with my waypoint... if my waypoint is indeed accurate.

 

Just for fun today on the way to work I marked a recognizable spot near the highway, with my Merigreen, I just checked the mark with GE, the location I marked was within 5-10 feet from where I actually made the mark, How did I come up with the estimate of accuracy? I stopped next to a Historical marker and took a picture of it after marking the location from within my vehicle, I got out of the Jeep walked around the front, went to the passenger door leaned my backside on the passenger door and took the picture. Since my jeep is about 5 feet wide and my GPSr bracket is mounted on the center console on the passenger side, which by my estimation was about 7-8 feet from the marker. The accuracy was not too bad I would say, of course your mileage will vary.

[+TrailGators]

Google earth is mostly designed/ slanted toward the "Oh Looky, there's my house" crowd. Don't bet on it's coordinate accuracy.

Google earth is good for giving you the lay of the land and that's about it.

[SiliconFiend]

That sounds similar to what I found. Google Earth seems to be five or so feet out of alignment with my waypoint... if my waypoint is indeed accurate.

Holy smokes, you're complaining about a 5 foot error? Your GPS will get you within 10 feet on a really good day (there's always the chance you'll get lucky and nail it right on). I've seen imagery alignment problems on the order of 50 feet, more in Google Maps than in Google Earth; for some weird reason they seem to have different imagery or alignment.

[+OpenTrackRacer]

I'm not complaining at all. I think it's pretty darn impressive. In fact, since it's within the probability of error from my GPS, Google Earth could be spot on.

 

That sounds similar to what I found. Google Earth seems to be five or so feet out of alignment with my waypoint... if my waypoint is indeed accurate.

Holy smokes, you're complaining about a 5 foot error? Your GPS will get you within 10 feet on a really good day (there's always the chance you'll get lucky and nail it right on). I've seen imagery alignment problems on the order of 50 feet, more in Google Maps than in Google Earth; for some weird reason they seem to have different imagery or alignment.

[+TrailGators]

That sounds similar to what I found. Google Earth seems to be five or so feet out of alignment with my waypoint... if my waypoint is indeed accurate.

Holy smokes, you're complaining about a 5 foot error? Your GPS will get you within 10 feet on a really good day (there's always the chance you'll get lucky and nail it right on). I've seen imagery alignment problems on the order of 50 feet, more in Google Maps than in Google Earth; for some weird reason they seem to have different imagery or alignment.

You are right, 10 feet off is excellent. I think it is alignment errors when they join the maps.

[robertlipe]

Coords loaded in Gogole Earth in appropriate datums are darned near 100% accurate. If you place a point and measure/mark it against a calibrated point, they'll usually match. What will mess with your head is imagery alignment.

 

It's fun to think that the plane/satellite is always exactly overheard when the shutter clicks and/or that all the imagery in Earth (or any similar program) is 100% orthorectified. But that's just not to be. The camera takes a foto and that foto has a bit of fish-eye action and that fish-eye action is warped to fit the curve of the Earth (which, inconveniently, is not only not flat, it bends the OTHER way) and is stitched to other fotos.

 

This is why that foto of the big bridge has waves (it was taken w/ multiple passes) and/or the coords say it's in the ninth tree when it's really in the tenth tree.

 

Imagery alignment problems happen. GPS error is real. Sometimes they cancel each other out, leading to the impression that all is perfect and sometimes they add with unfrotunate phasing, giving te impression that things are really bad.

[oxothuk]

Like others have said already, the "inaccuracy" in Google Earth comes from misalignment of the satellite photos. They have to line up a photo which has some perspective distortion against a map of the earth, which also is not flat. I suspect (don't know) that they may fix the image on known landmarks when possible.

 

Quite often if you turn the display of roads on and off you can see that the roads don't line up exactly with the visible image of those same roads in the satellite photo. It's usually pretty close, but not exact.

[table/cache]

I think it's telling that the US/Canada border is not aligned with obvious border features in my area.

 

The pointer readout places it to the South, and the line on the "map" places it to the North, whereas the image is in between those 2.

Edited April 10, 2008 by table/cache