I need to obtain ground truth data of Google Earth for a certain region for the purpose of constructing confusion matrix. The data should correspond to Landcover classes.

My main aim is to construct a confusion matrix with rows and columns classified as land covers. Now I already have a experimental data set of latitudes and longitudes corresponding to each land cover classes. So I suppose I have my rows are filled. Now to obtain column values I need to fill it with ground truth data of Google Earth. However I am not exactly sure what does it mean in this scenario. I am guessing that I need to enter the lat-long values in GE and obtain land cover classes from it corresponding to the imported lat-long values.

Am I right? If yes then how do I get the land cover classes of the imported lat-long regions?

  • 1
    What does "Ground truth data of Google Earth" mean? I thought "ground truth" meant the truth on the ground? Google Earth has no "ground truth" - only the ground has that. And does it even have land cover data?
    – Spacedman
    Commented Feb 12, 2017 at 17:51
  • @Spacedman Sorry! Maybe I wasn't much clear in my question. By ground truth data I mean I have to look at all the lat-long coordinates and decide for myself the land cover category to which they belong. I have more than 2000 data points. So instead of looking manually, is there any other way of extracting this land cover data? Commented Feb 13, 2017 at 10:02
  • Can't you find land cover data for your are of interest? I'm pretty sure there's land cover data for Europe available. Anything you try and derive from Google Earth will need some processing and will be subject to error as well, since its just satellite imagery.
    – Spacedman
    Commented Feb 13, 2017 at 10:29
  • @Spacedman Yes! But that isn't my project. I need to use the Google Earth in order to do ground truthing. Regarding errors, every land cover data will have some sought of error. That's why I'm trying to construct a Confusion matrix to determine GE's accuracy (and Simultaneously of my data). Commented Feb 13, 2017 at 13:07
  • Why do you need to use GE? Are you saying you can't use anything else, even a perfectly good land cover data set? If so then you should probably open a new question "Deriving land cover classes from Google Earth imagery"
    – Spacedman
    Commented Feb 13, 2017 at 14:07

1 Answer 1


Building a confusion matrix takes time and this time should be planned in any map production project. The confusion matrix is obtained by counting the number of occurrences of the pairs "map class/real classes" for a set of (usually) randomly distributed coordinate points. You can extract the class of the map for each sampling point using basic GIS tools. However, the only way to extract a trustable reference dataset on Google Earth is by photointerpretation. The best way, unfortunately rarely done in practice due to the cost, by going on the ground. Of course, even if the photointerpretation is not perfect and therefore can get some help from ancillary data if you are unsure (especially if you do not know the area that you are validating). However, at least you can avoid the bias of another classification. It is futhermore recommanded to use a higher resolution image for the validation than your actual product, and this also mean that your photointerpretation must take the different scales into account.

In practice, Google Earth and other services that provide very high resolution images are accepted for the validation of land cover maps. As you said, you could enter the Lat/long coordinates of each of your points and write theire photo-interpreted land cover to create you validation dataset. Idelally, i would seek a system to directly write the values using a drop-down menu to go faster. At least, import your spreadsheet to GE in order to make sure that you have the exact location of each point.

Alternatively you can compare your map with another map, but you will not know if the conflicts are due to errors in your map or in your reference map.

as a final remarks, some great initiatives are distributing their validation points. For example the data from "collect earth" (soon to be released) could be very usefull in your case.

  • 1
    How can one extract a trustable reference dataset on Google Earth by going on the ground?
    – Rim Sleimi
    Commented Jun 18, 2020 at 17:21
  • Thank you for your remark, my sentence was not correct. I meant: Google Earth photointerpretation is not as trustable as going on the ground, but it is accepted as a source of "trustable" reference dataset when going on the ground is not an option (e.g. unreachable area, very large study area...). Of course, this data will not be perfect but what is important is that 1) it is substantially better than the data that you validate and 2) it is independant of the map production process.
    – radouxju
    Commented Jun 19, 2020 at 7:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.