Is Google Maps more accurate than a GPS?

[+harryatkins]

I wondered if anyone had any views on the IPhone aGPS and Google Map being used for finding / marking caches, rather than actual GPS devices.

 

I do a fair bit of geocaching to keep my skill up with my MAP60csx.

 

A few times now, after I have marked a location (letting it average from over 2 mins) and getting an accuracy to under 10 metres, someone has come back to me, which seems to be fair and something I want to try and research.

 

"Just a little advice though, have a look at your coords on google maps, I'm assuming you know how to enter lat/long into it, as they don't map to where your caches are. The nano actually maps to the pathway alongside the church, The final cord’s also map into the field rather than the trees, I think we were indicating 12-15m from each cache when we found them"

 

What some seem to be doing is using their IPhone and a geocaching app http://www.geocaching.com/iphone/ hooked into Google Map to see where the cache is located. This seems fine (although has the obvious restrictions) but there always seems to be some discrepancy when you look at the coord from the MAP60csx in Google Maps / against ground locations.

 

I know that the IPhone 3G uses aGPS (assisted GPS), which means that the cellular network aids in positioning the handset, giving the GPS hardware less number-crunching to deal with. That means you get super-fast GPS lock-on times. But, it seems that some iPhones are getting conflicting data from the 3G data network and their built-in GPS receivers, leading to incorrect GPS positioning. The accuracy has been noted as 50+ foot, which I suspect would give you a of target coord if then linked into Google Maps.

 

You can see that situation.

 

A chap in a mountain rescue predicament calls in and gives his cords taken from Google Maps (which is in decimal format). If the mountain rescue team then enter the details into the GPS devices which I am sure they will use rather than and iPhone there could be a conflict and maybe danger.

 

So if I use Google Maps to obtain a Lat and Long coord and put it into my GPS device would I get to the expected location. The same said vice versa. If I was out and took a mark on my GPS devices and noted the coords and then put into Google Maps would I expect to be "pinned" at the same location as where the mark was made. In my experience the answer from the later example is no. The mark taken on my GPS which is a MAP60CSX (Garmin) did not relate to the position I was on the ground when I viewed in Google Maps. I am trying to work out why there is such a discrepancy.

 

I believe my device to be calibrated ok.

 

Appreciate any thoughts that you have on this subject.

[+keehotee]

Let's start with an obvious comparison...

 

Have you tried getting the co-ords of an obvious landmark from Google, loaded it on your GPS, then navigating to it using your GPS?

Or comparing Google co-ords with paper map co-ords?

[+user13371]

RE Google: It depends where you're looking. In some places, Google's imagery is perfectly referenced --- and in some places its very far off.

 

If you want to check Google's images for your area, use your trusty Gamrin to set some waypoints on spots you can easily identify later online -- and then look at those coords in Google Earth to see how well they line up.

 

RE iPhone versus Garmin (or any other dedicated GPS): The iPhone's GPS is better suited for finding cafes than caches. Under ideal conditions and with 3G turned off, it is certainly good enough for geocaching. And being able to see your cache coords on an aerial photo helps a lot even if the GPS fix isn't perfect. But the ocassions where it actually puts you in arms reach of your real location are rare.

 

I usually get to a cache with my iPhone. And if I can't find it, then I take my "real" GPS out of other pocket to zero in.

 

Hope that helps.

Edited December 17, 2009 by lee_rimar

[+twolpert]

Even if GPSrs of all types were created equal, you still have deal with the inherent limitations of the system and differences in conditions between the time you hid it and the time they looked for it. With consumer-grade hardware and without post-processing, 3-5 meters is about the best you're going to do on the average. If you're off by 5 meters when you hide it and the finder is off by 5 meters in the other direction when he looks for it, you've got a 10 meter difference. And that's without considering the possibility of larger errors due to the possibility of bad satellite geometry on one or both occasions.

 

My take on averaging is that it requires a lot of samples to do any good (5 minutes or more). And that it's only good for correcting things like multi-path effects which have a random component. If you have bad geometry and a high EPE over the time you average, it's not going to help. The average of bad data is still a bad position. If I am hiding a micro or something else which will be hard to find, I revisit the cache on several different days and average the results. Particularly if I pay attention to the constellation, this helps eliminate errors due to high EPE.

 

Also, want to echo Lee's comments about Goolge Maps or Google Earth. Most times, the imagery is properly georeferenced. But there are cases where it is not. You can't put any faith in it until you check out the georeferencing in a particular area.

[+harryatkins]

Guys

Thanks for the quick responces to this one, there is already some good suggestions that I will be looking at plus some interesting background stuff.

Harry

 

Also picked this up from the new service on Aardvark.

Google maps are based on the USGS survey which determines the actual

lat/lon coordinates on the map. The lat/lon system was determined long

before GPS was invented. So, Google maps is showing the exact

predetermined lat/lon of any position on the map.

A portable GPS uses satellite triangulation to try to determine it's

location on the earth or in the atmosphere in relation the

predetermined lat/lon coordinates on the map. GPS is good but not

perfect.

Provided by http://www.maps-gps-info.com:

GPS accuracy is affected by a number of factors, including

satellite positions, noise in the radio signal, atmospheric conditions,

and natural barriers to the signal. Noise can create an error between 1

to 10 meters and results from static or interference from something

near the receiver or something on the same frequency. Clouds and other

atmospheric phenomena, and objects such a mountains or buildings

between the satellite and the receiver can also produce error,

sometimes up to 30 meters. The most accurate determination of position

occurs when the satellite and receiver have a clear view of each other

and no other objects interfere."

Edited December 17, 2009 by harryatkins

[+Minimike2]

I would believe my PN-40 over Google and iPhone, day in and day out.

[+tec_64]

Up here in Sudbury Ontario, I've used google maps satellite photos via the AcmeMapper website to get fixes on various precision landmarks. Punched them in to my Garmin. Used Trimble planning to pick a day and time with low HDOP and have got pretty decent (read accurate) and repeatable results.

 

Wayne

tec_64

[Ken in Regina]

I've been using a really interesting Google Maps mashup that helps me turn my GPS handheld into a Golf GPS device. I also have a golf GPS application for my iQue 3600 (Palm PDA with integrated GPS) that lets me take coordinates from Google Maps and type them into the application.

 

Both of these let me use the aerial views of Google Maps to set the locations of greens and various hazards on the golf course.

 

In doing this, I have used my own GPS devices to check the locations that I've input from Google Maps against the locations of these things on the golf course. I've found that the error between my devices and the Google imagery is not huge, probably not enough to mess up a search and rescue operation. But it's enough to screw up your golf game in most cases. On the local golf courses the error is anywhere from 30 ft to 100 ft.

 

...ken...

[timz2]

If you want to know how accurate Google Maps (or Bing) is in your area, just go to

http://www.ngs.noaa.gov/cgi-bin/ds_radius.prl and get a bunch of known locations in your area. Water tanks, churches, radio towers-- or discs, if you can go there and see where it is in relationship to something that's visible in the aerial pic. You'll be surprised how good Google and Bing are.

 

Naturally you use the center of the base of the water tank, not the top. You'd think we couldn't expect them to do well with landmarks on the tops of hills, but seems like they do better than you'd expect there too.

[+edscott]

I've found Google maps in most areas are more accurate than a GPS, but since the hider used a GPS to obtain the coordinates you are stuck with their numbers regardless of the accuracy. If the real question is whether you can use Google to find those caches then yes, from my experience it's quite doable.

[mtbikernate]

If you used Google Maps alone to search for my cache, you'd end up on the wrong street. Google Maps streets are just flat out wrong. There are 3 dead end streets off of the neighborhood for the cache. The cache is on the southernmost street. But Google Maps says it's on the middle street.

 

The problem is that they have a street marked where there is none, and it throws everything else off. If you look at imagery alone, you'll see. But if you look at the street or hybrid views, it's off.

 

How it's more accurate than the GPS, I don't know. The GPS coordinates show up where they're supposed to on the imagery. A GOOD mapping company will georeference using actual ground-truthing with a survey grade GPS. That's going to give them better accuracy than your handheld. Chances are, they use subfoot receivers with differential correction, and possibly centimeter accuracy receivers with differential corrections. That will allow the imagery to be georeferenced to an extremely high degree of accuracy.

 

But, in some areas like mine, they're not going to ground truth everything. Close enough will make them happy. So, there will be problems and errors like the one I described. It's obvious to me that the maps are not consistently accurate.

[+Prime Suspect]

If you want to know how accurate Google Maps (or Bing) is in your area, just go to

http://www.ngs.noaa.gov/cgi-bin/ds_radius.prl and get a bunch of known locations in your area. Water tanks, churches, radio towers-- or discs, if you can go there and see where it is in relationship to something that's visible in the aerial pic. You'll be surprised how good Google and Bing are.

You shouldn't be surprised at this. What do you think they use to line up the aerial photos? It's when you're not near these landmarks that the errors crop up.

[+StarBrand]

Virtually all published maps have an error built in that varies a bit with location. Most printed maps have an error of about 90 feet at well known locations. Some get as low as 40 feet at well known locations. My local experience with Google maps is that it runs from agreeing with my GPS units to being upwards of 150 feet off just a few miles away. Given that variabilty, I'll trust my GPSr every time.

[timz2]

What do you think they use to line up the aerial photos? It's when you're not near these landmarks that the errors crop up.

So when you're within... say, 5 km of a surveyed landmark, you can expect Google/Bing to be correct within... how many meters?

[mtbikernate]

What do you think they use to line up the aerial photos? It's when you're not near these landmarks that the errors crop up.

So when you're within... say, 5 km of a surveyed landmark, you can expect Google/Bing to be correct within... how many meters?

 

It's not that consistent. It depends on how the image was orthorectified (which algorithm).

[Ken in Regina]

So when you're within... say, 5 km of a surveyed landmark, you can expect Google/Bing to be correct within... how many meters?

That's way too far away to say with any certainty. Much too far.

 

Just to illustrate, as I mentioned above, I've done a lot of messing around with my eTrex Legend HCx to use it as a golf GPS. I have found a fairly significant variation in accuracy of Google Maps versus the GPS reading from one part of a golf course to another.

 

If you are as much as 5 km away from a spot they used for calibration, the error could be nearly anything. Mostly it would be unpredictable. But I doubt you'll find many areas where you would be as much as 5 km away from a calibration spot.

 

...ken...

[timz2]

What do you think they use to line up the aerial photos?

Maybe so-- I certainly don't have a better explanation of how they do it. But:

 

KU3867 is alleged to be at 40.787750 N 73.810427 W; if you go to that lat-lon on Google/Bing you'll end up maybe 20 meters south of the church dome. I've never confirmed it, but it's a good bet it's the church that's off, not G/B; the present building isn't the one C&GS triangulated. So: how did they know to ignore this landmark?

 

KV4906 is alleged to be 40.788858 N 74.254382 W; this time we know the present tank is not at that position. Maybe G/B saw the recovery, but looks doubtful.

 

LX4067 was 41.100482 N 73.771478 W; the tank has been removed. G/B knew to ignore the nearby new tanks.

 

KV4430 was 40.641898 N 74.078254 W; think the 2001 recovery is why G/B doesn't line up with the new steeple?

 

KU3273 was 40.684800 N 73.627982 W; maybe the recovery warned G/B to ignore the new tank?

 

HT2797 was 37.716096 N 122.395055 W-- the foundation seems to still be there. Did the recovery warn G/B to not use the new tower?

 

HT2827 is the most recent change; the old tank was 37.723034 N 122.424366 W and some of the Bing pics show the tank there, but Google shows the new tank and ignores it.

[+Prime Suspect]

You're assuming that NGS Benchmarks are being used to align the maps.