2

For a piece of coursework we have an aster layer that captures the tops of trees and a panorama layer that sees ground height (both are raster) so we have subtracted one from the other to get tree height but the values were are getting are the difference between pixel values rather than height values. Can this be changed to a value in metres?

  • 1
    Surely the result of your subtraction will be in the same units as your inputs. If not please clarify how/why you are doing the subtraction so that the output is different. Also, I urge a note of caution in this process as you might end up with negative tree heights or unrealistically tall trees (depending on the resolutions, sampling methods and errors of the respective inputs). You will need to have a lot of sanity checks on your output. – MappaGnosis Dec 2 '14 at 11:59
  • 1
    @Chloe What do you mean by 'pixel values'. I assume that the two input DEMs contain elevation values (both in metres?). Subtracting these raster could only result in an output of difference in elevation, also in metres. As MappaGnosis correctly points out, depending on the data source and resolution of the inputs, you may end up with some rather unrealistic elevation differences. You should also check to ensure that the same units are used in each of the input DEMs. – WhiteboxDev Dec 2 '14 at 12:27
  • @mappagnosis This is exactly what has happened! We have very unrealistic values but this is the way we have been advised to work out the height difference. Why is it like this? – Chloe Dec 2 '14 at 12:38
  • 1
    Your two height data sets probably come from different sources and have independent vertical and horizontal errors. If the vegetation was automatically sampled (e.g. LIDAR) the sample point may hit the earth between trees. So, for example, let's say that at an identical location the bald earth DTM is over-reading by 2m and the LIDAR hits bare-ground and is under reading by 2m, you get a tree height of -4m. When you factor in horizontal error and mismatch of sampling resolutions, you can compound small individual errors even more dramatically. What are the resolutions of your data? – MappaGnosis Dec 2 '14 at 14:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.