I noticed today that there is a discrepancy between the map and satellite views. This can be seen by observing a recognizable structure (e.g., a house) as shown in satellite view and as shown on the map view. I am wondering which view is considered more accurate. The differences are not large - in the case of my house I would estimate the locations to differ by about 30 feet with the map view location being about 25 feet west of the satellite view location. I have noticed this only once - today - and am wondering if this is a regularly occurring phenomenon. The map and satellite views are both layers based on the same GIS coordinates as collected by the camera and those coordinates are also suspect.
1 Like
Deal with this a lot in OpenStreetMap, in my experience Maxar imagery is generally pretty inaccurate and that tends to be what Google uses. I generally use ESRI imagery when editing OSM and its significantly more accurate then anything Google has. I typically just use the coordinates off my camera and trust that it is right even if googles aerial imagery makes it look like it is off.
This is a pretty common problem working with base maps in GIS. It’s easy to take for granted that Google (for example) is accurate, but that’s far from the case. I first realised this when I found myself involved in a project located on the join between two “units” of Google imagery where dealing with the discrepancy posed major problems. Unless in an area with very bad satellite coverage, I too tend to go with my GPS readings and not worry too much about their location on the satellite or topographic map. Unless of course you are producing graphics, in which case a bit of “manipulation” may save the day.
1 Like
As bryannamar and lynkos said, aligning maps is a common issue in GIS. There are a number of reasons this happens; one is projecting a flat image taken from a satellite onto the spheroid model of the Earth that underlies the GPS data. Some of the sattelite image may be perfectly aligned, but in areas further from where the pixels have any actual georeference, distortion becomes greater.
1 Like
complex terrain with a lot of relief can complicate that rectification process, too.
also large areas with few really good points spread across the image to use as references for the rectification process.
I did my share of Landsat image rectification for my master’s thesis research. one of my most hated GIS tasks.
This is a rather curious, but well-known question. The point is that, for historical reasons, Google Maps and Google Earth use different coordinate systems. When you use the satellite view, the Earth is shaped like a WGS 84 ellipsoid (EPSG:4326 - https://epsg.io/4326). But. In map view, the Earth becomes (suddenly) spherical (EPSG:3857 - https://epsg.io/3857) :-). Of course, such a “change” in the shape of the Earth causes a shift in coordinates on the ground (small, in fact). Certainly, a more precise model is the WGS 84 ellipsoid. The Earth is not round!
A little more details: https://gis.stackexchange.com/questions/48949/epsg-3857-or-4326-for-googlemaps-openstreetmap-and-leaflet
3 Likes
In addition to all the reasons others have given, Google Earth uses a mix of satellite images and any given satellite is in a different position. As a result the angle of the image can be very different from image to image, and in complex terrain, near islands, or anywhere there is something tall and steep the peak can wind up being offset by quite a bit, and one side of the feature completely hidden.
This is a slightly different (but related) thing than the rectification issue.
If you open Google Earth Pro and click the clock icon you can scroll back through images taken at previous dates and see how much the satellite imagery and specific locations move around.
I work in a really complex karst landscape and this is actually a bit of a problem for us, even with high end imagery and DEM data.
1 Like
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.