How accurate are Google Earth images
The issue of the accuracy of satellite and orthorectified images of Google Earth is a record question in the search engines, in these days in which confusing precision with tolerance is as easy as losing the GPS in the taxi, it would be good to make a couple of analyzes about the convenience or not of using this data for serious work.
Recientemente DigitalGlobe, A provider of Google Earth, Announced that they will soon have more material after the launch of a new satellite of higher resolution and greater coverage per day.
This satellite is called Worldview I, Which will work in conjunction with a Quickbird satellite, the images will have a pixel resolution of 50 centimeters (currently 1 meter) and will be able to capture 600,000 square kilometers per day, which they currently get in a week.
This often misleads those who only concentrate on the final product (maps) and not on the intermediate products of survey and its photogrammetric origin which is what allows us not only to have products but also their level of precision and pertinence. When using Google maps material For local purposes It can be corroborated by the lack of precision, sometimes up to 30 meters, since a satellite image requires a dense network of first order and local checks for its rectification. Not that it's bad, it's for the purpose of Geographical web which is the way Google Earth calls that layer.
For those who still want to do cadastre with ortos of GoogleEarth, here is an example of what will be found:
This is La Jaguita, Comayagua, Honduras; Observe how you see in the splicing of images the repetition of the same complex of buildings, 36.51 meters to the northeast.
We use two GPS Magelan Mobile Mapper, one as a base, another to raise points and after making the differential correction, we compare it with an orthophoto created with aerial photography of three years ago with 5,000 flight; the GPS data matches perfectly with the orthophoto but not with any of the Google Earth images, one of them is 23 meters to the northeast, the other to 19 meters, but the displacement vector has a different angle.
In this respect it is not possible for us to question a system, which is the best that many of our towns can have, where cartographic information will not arrive in the next 10 years, and to top it all is free; What is important is that we are aware of its relevance and limitations.
To meditate I leave you this point of the terms of Google Earth, the one that nobody reads and only presses the "I accept" button:
E) NO ADVICE OR INFORMATION, WHETHER ORAL OR WRITTEN, OBTAINED FROM GOOGLE OR THIRD PARTIES OR THROUGH THE USE OF THE SOFTWARE, GRANT ANY WARRANTY THAT HAS NOT BEEN EXPRESSLY SET FORTH IN THESE TERMS AND CONDITIONS.
I hope the experience will be useful to someone: I surveyed 14 hectares with a gpsmap 64sx, with an antenna, this time the coordinates were around 1.5 meters of precision in planimetry (not recommended for altimetry), to have a more precise data I measured the boundaries (not all because there were swampy sections) and link arbitrary points with measurements and adjust the coordinates in ch. thus the margins of error are reduced to centimeters, I cannot explain the procedure in detail as it would be very tedious, I hope it helps you.
Note: the GPS navigator type in topography is used to have coordinates approx.
RTk gps are specialized devices, with a millimetric margin of error.
I tell you colleagues Google Earth has a lot of error and more at heights, I was raising with station and the margin of error of this regarding my heights was almost 8 meters in a project where a meter of more or less could cause the aqueduct to fail in a relatively flat area, rest when you see all these comments thanks.
They will never match you. The model of elevations of Google Earth is very simplified because it is for the purposes of a global terrestrial globe; which does not happen when you do a local survey with more altitudinal control points.
I commented. I just did a topographic survey for drinking water system in Arequipa and the levels that I have obtained with the Total Station do not coincide at all with the levels or contours of google earth. There is a difference of 40 mts of altitude.
It is necessary, define that many confuses GPS sub-metrics with browsers (garmin, Magellan, etc) no doubt for there are Sub-metrics with cadastral precision of 5 cm (which is relatively large) to browsers that have differences of 3- 5-10 meters.
Google images should be taken as a reference, but not as a basis for performing sub-metric work; Because and until they confirm it, the variations will continue with the different devices that each one uses.
For Mexico the error is from 5 to 7 meters, I have already carried out several tests with 2 GPS. However, when you use the "Geolocation" function in Autocad, this error decreases taking into account only the GPS error, according to the tests I have carried out so far.
It is true a survey made with total station does not agree at all with the images of google,
You should check what sam you are working, usually working on sam 84 is used internationally, because that makes the difference of a point.
They are from 7 to 10 mts app, I can tell you with a base because I use the gps and military letters a lot because of my work and it denotes the differences.
today I found myself with the same doubt and was looking for the highest precision for reporting data, and the difference in data is very wide that really makes me distrust a lot. the average difference with gps and google is more than 5 km and that is worrisome
DO NOT USE IT FOR PRECISE DATA
The DATUM of UTM coordinates used by Google Earth is WGS84, and the difference will be due to the GPS you use being configured in the NAD 27 Datum. Configure your GPS, I recommend it. Greetings from Managua, Nicaragua.
Well, that's not much that we say in precision. Not for measurement of plots, unless the use is only fiscal or does not link a legal utility.
Hello, very interesting comments. We appreciate the free Google Earth, however I think that if someone requires more precision, without investing large amounts of money, I could use a GPS, which provides a precision of more or less three meters. Please someone to tell me if my comment is true, since I often require accuracy in locating points. First of all, Thanks.
When I started using Google Earth and GPS I had the same problem, the solution is that both must be in the same coordinate system.
For that you will be difficult to use the image of Google. Depending on the country and the area where you are, the trips are different.
Perhaps one way is to lower the image with Plex.Earth, and take control points to more or less rectify it to acceptable precisions.
I am dedicated to navigating in 4 × 4 vehicles and we constantly use the Garmin gps, I wanted to mix it with Google coordinates and I have differences of up to 200 meters between the image and the exact point so that affects me a lot to locate the gaps and of course the waypoint, how could I make this square or what programs do I use with satellite photographs to have the least variation? I want to handle with data, photos and coordinates in real time.
Yes, this link speaks about it
Anyone know if there is any improvement in these versions?
First of all thank you for the answer, I dipped a bit of the information you present in your various publications about cadastre, I think it will be somewhat complicated to use Google Maps for cadastre use so I go to find other solutions, if someone here know some other alternatives I would be grateful as I am looking for some solution to address a problem that goes hand in hand with cadastre.
That depends on the geographical location. For example, in the United States and some European countries the image you see in Google Earth has been provided by the states or governments. So there the precision is very good.
But in other countries, precision walks between 10 and 30 meters scrolling images. That can be corroborated in the splicing between shots of different years.
There is no paid version that gives you better images.
Hello good day.
I'm interested in using google maps, for different projects with different expectations, so I'm interested to know how accurate it is or what its error rate is; I am also interested to know if there is a version of pay that better accuracy or simplemte where I can inform of this.
Thank you very much, thank you for the publication.
Hello, in my opinion I do not think that the geocentric coordinates (k) should be corrected, as I read there, but rather ..., the displacement due to the relief that is produced by the parallax effect.
Download it with Stitchmaps as a jpg at a height of no more than 500 mts, with its calibration file and then uploading it from Ilwis.
If Ilwis does not read the calibration file you will need checkpoints with which to reference it.
How can I pass a google eart image to the 3.2 ilwis program and give me good foresight as to the coordinates.
The article is very good, in this regard… the fact that it is not precise does not make it bad and as they say it is the only place where they offer us “free” images of such a high resolution. I have used some and I have been able to verify that the error increases in mountainous areas or with uneven topography, but in flat areas the error compared to GPS navigator data has not been very relevant... logically it depends a lot on the purpose of use, I consider that for “local” cartography it is an excellent tool that saves a lot of time and effort…always safeguarding the support with many ground control points for its rectification.
Just as my email says, one thing is flat topography and another is topography from space (which does not exist ...) but photogrammetry from space must correct geocentric coordinates based on the curvature of the earth, see the difference TOPOGRAPHY FLAT …… CURVATURE OF THE EARTH… in addition to that, the margin of error of the lens with which the image was taken from space, all lens is not flat 100% is curved, so the fish eye lens, 50 mm 200 mm or zoom De This explanation is where the error of the above image is derived and whoever tries to do a cadastre with google is cheating the owner or government
The problem is that the phase shift is not uniform, it is not only moving them a certain distance because it is fixed at one point and then decomposes at another. I have used them, downloading them with stitchmaps, then rectifying them with more control points using Microstation Descartes ... and this improves them a lot ...
Google plans to integrate geoeye images… but to date, satellite images do not have good georeference unless they warp with enough local control points.
I do not know of another server of better quality (by google), yes there are services OGC of some countries that offer own material in the form of web map services.
Instead of having nothing, it's not bad. For cadastres with a tax approach, land use…. it is fine but not for cadastre with a legal approach, which will include land titling…. it would be difficult to give a property to someone, and that it is 30 meters in the title of the neighbor
HELLO FIRST A GREETING WHAT INFORMATION DO YOU HAVE ABOUT THE NEW GOOGLE PROJECT… I UNDERSTAND THAT THEY ARE GOING TO TAKE A NEW NETWORK OF PHOTOGRAPHS. LOOK MY QUESTION IS FOR THE FOLLOWING I AM ELAVORING A CATASTRAL PROGRAM AND I WORK A VERY GREAT WORK IN ADVANCE USE THE PICTURES FROM GOOGLE EARTH AND HOW YOU DO THE COMMENT THEY HAVE A DEFACE OF A FEW METERS IT CAN BE MODIFIED SO THAT IT EMBODIES WITH WHAT I NEED BUT I WOULD LOVE TO KNOW IF THERE IS ANY OTHER SERVER WITH BETTER QUALITY
OF ANIMATION A GREETING
IN CONCLUSION, THE GOOGLE EARTH IMAGES DO NOT SERVE ??
WHAT DARIAN RECOMMENDATION
GIS TRAINING JOL
Yes, they have already been Shown Pixel images of 41 cm.
I wonder if by now be running the new satellite?
Congratulations, for your comments, always interesting.