I have been given a figure by a non-surveyor derived from google earth, of the volume of several piles of material.
I have good survey data of the ground before, and my own survey data of the material piles. I would suspect that I could have done a denser topo, but I believe what I have is adequate for relatively uniform piles of material.
In your experience, how accurate is google earth?
thanks in advance as always.
What is the terrain source of the google data? USGS 30m DEM? 10m? 3m?
Local photogrammetric contours? LiDAR?
DDSM
dont know - the trucking company says they have a figure derived from Google Earth -
Assuming its the best data available, what kind of competition do I have?
Certainly not survey grade. Look up my address and you will be a couple hundred feet away. Look up my Dad's house address and you'll be pointed to the local town square. That would be several miles from his house.
i would suspect that they merely just did a toe to toe volume, wouldn't really take notice of whatever they have if its from google earth
> I have been given a figure by a non-surveyor derived from google earth, of the volume of several piles of material.
>
> I have good survey data of the ground before, and my own survey data of the material piles. I would suspect that I could have done a denser topo, but I believe what I have is adequate for relatively uniform piles of material.
>
> In your experience, how accurate is google earth?
>
> thanks in advance as always.
You could ask who exactly is standing behind the volumes quoted "from google earth". You are standing behind your answer. The chances of getting someone from Google to stand behind their answer? Zero, no chance at all.
If they still balk at that logic you can tell them I said the volume is 6 ... the answer is always 6. (Notice I didn't say what units or how I arrived at the answer.)
Larry P
> In your experience, how accurate is google earth?
I've found terrain data (and aerial imagery) in Google Earth to be wildly inconsistent. I'll leave this KMZ example on-line for a while. Within this relatively small area, the difference between the elevations shot with RTK GPS (with Geoid 2012a model applied) to that of Google Earth varied between -9.8' (deep red) and 2.2' (light blue)... a range of 13' of elevation difference (for a browser-based map, visit http://goo.gl/maps/eZNWT ). Some elevations are very close (the areas close to a shade of white); others, not so much. I personally would not put a lot of faith in a volume report from a surface that comes from a consumer-grade application.
Honestly, I would just wish them good luck, and make sure they have my number for when the crap hits the fan. And it will, when the billing doesn't work right for the trucking.
It's like the call you get that starts out.. "I really don't need a survey". Ok, then hang up because I have a line of people that do need me waiting.
Next!
Andy
> If they still balk at that logic you can tell them I said the volume is 6 ... the answer is always 6. (Notice I didn't say what units or how I arrived at the answer.)
I beg to differ. The answer is 42. It's always 42. 42 is the answer to life, the universe and everything.
Interesting - Could you please post the RTK data (ASCII) that you used to make the comparison. In the area of good point density and less brush at the bottom of slope, Google Earth appears to match fairly well. On the side slope different story. How do points compare on a one by one check? Is there much height variation between the shots on the "M" and the ground immediately next to it?
About the original question, I would not be too quick in disregarding the results obtained with Google earth. The absolute accuracy might be out but the relative could be surprising.
My questions would be:
What size piles are talking about here?
What software was used to determine the volumes of the GE imagery?
Don't get me wrong, I wouldn't use Google Earth for volume calcs, unless an astronomic size of pile is to be measured where the possible accuracy shortfalls of Google Earth could be diluted in the error budget. Back to the original question, my answer is "I do not know".
all the responses have been very helpful. I do very little with google earth so I didn't want to make assumptions without going here first. I know nothing about the google earth data. I also do not want to be more specific in case someone involved with the project is able to figure out I am talking about their project on a global discussion board.
The size of both volume calculations are well less than 10,000 cubic yards. Its the relatively small size and significant difference between the two that first gave me great suspect about the google earth. My own survey would have to be way, way, way off to be that bad.
> I have been given a figure by a non-surveyor derived from google earth, of the volume of several piles of material.
>
> I have good survey data of the ground before, and my own survey data of the material piles. I would suspect that I could have done a denser topo, but I believe what I have is adequate for relatively uniform piles of material.
>
> In your experience, how accurate is google earth?
>
> thanks in advance as always.
Am I missing something? If you have information on the same ground and piles, wouldn't you be able to answer that question for us?
My best guess from playing around with the historic imagery in google earth is that the "elevations" are from some form of quad sheet data. I have a pit that has been mined by our company since 1994 and the elevations remain the same approximately as what you might expect to interpolate from the quad sheet. Unless those piles have been there longer than the latest quad update, I would not trust them.
:good:
LOL, Great movie. I once attended an eminent domain cont. ed. class. He had an opening portion that set the tone for the class and made it very interesting. It was clips of videos showing how Hollywood has villainized eminent domain. Everything from Andy Griffith to the James Boys.
After the class I asked if he liked sci-fi, because he missed the best dual-reference of all: Hitchhiker's Guide to the Galaxy. It's even in something like the first 15 minutes or so...
>
> Am I missing something? If you have information on the same ground and piles, wouldn't you be able to answer that question for us?
I'm just guessing that Borderline Survey Pro has been told by a client that they have a better answer that proves his survey is off and way more expensive than he has any right to charge.
Sounds like the pile was pretty uniform. A simple figure calculation should provide you with a good check.
The two answers (simple figure math vs. survey) should match within reason if the pile was uniform. Good thing to have in the file anyway.
I am pretty sure that Google Earth does the same thing that Google Search engine does. It crawls the web indexing sources of available data. I know several states make DEM data available at varying degrees of accuracy on web pages. Google just takes what people put out there for free and adds it to their "product". It may or may not even get a QA/QC. I know there used to be links where the rectification was off and you could see a road out west somewhere that didn't meet from one tile to the other.
Lately I have been seeing consistently well under a meter when I use Carlson or TBC to export to Google Earth in urban areas.
Google Earth elevatins are hit and miss...and by hit I mean the position seems to match my elevation within 2 feet. By miss I mean several hundred feet off vertically. I wouldn't use the google earth vertical components for anything more than initial planning.
> Interesting - Could you please post the RTK data (ASCII) that you used to make the comparison. In the area of good point density and less brush at the bottom of slope, Google Earth appears to match fairly well. On the side slope different story. How do points compare on a one by one check? Is there much height variation between the shots on the "M" and the ground immediately next to it?
All of the source data for that project site is at:
http://update.carlsonsw.com/user_conf2013/Elective/Elective%20--%20Session%208.zip
I should mention I took some liberties while collecting that "M" data back in 2009. Namely, I was about 25 miles outside of the Iowa Real-time Network and collected the site without localizing and without a regional geoid file. At the time, the WisCORS network did not have coverage in the area and I didn't have a base GPS available shorter baseline corrections. I collected it merely for field-to-finish coding demonstration purposes. I've since re-processed the source GPS data with the 2012a model to try and remove that aspect of the elevation difference.
Since the terrain around the "M" site isn't a typical for construction or stock-pile purposes, I did another example this morning using available Iowa LIDAR data around the Field of Dreams movie site (figuring it'd be flatter and is in an area where I know LIDAR data exists to remove the "human" variable). http://dl.dropbox.com/u/30460571/GESurface-LIDAR.km z">These results are even more pronounced with a maximum cut difference of 11.7' and a maximum fill difference of 8.2' (red = cut, blue = fill) and I'll leave the KMZ File file on-line for a bit.
The source data was pulled from http://www.geotree.uni.edu/lidar/ and converted from UTM Metric to IA North US Feet.
The information provided by Google Earth is remarkable considering its world-wide coverage but I still wouldn't trust it for volume calculation purposes.