Loading ...
Sorry, an error occurred while loading the content.
 

AI-GEOSTATS: Again: Variograms or Kriging for Digital Elevation Models Analysis and regularization ?

Expand Messages
  • Rainer Mueller
    Dear AI-Geostats members,thanks for your reply ! ... distribution ... that ... I already looked in detail at the distribution of height differences and found a
    Message 1 of 1 , Mar 29, 2001
      Dear AI-Geostats members,thanks for your reply !

      > An interesting question. You can, of course look at the entire
      distribution
      > of errors, rather than just the summary stats. I would recommend doing
      that
      > just to know wether the error is normally distributed for example.

      I already looked in detail at the distribution of height differences and
      found a gaussian distribution, but shifted with an offset. This is ok,
      because I compare a Digital Surface Model (optical data) with a terrain
      model. I also separated my DEM heights into objectclass-related heights with
      a classification layer, because i.e. imaging and matching trees will give
      the height of the trees, rather than the terrain height. But I only found a
      trend for expected 'tree' and 'building' height offsets (speaking of
      comparing it with a digital terrain reference model).
      What I want to know is: Is it possible to find something like the 'level of
      information content' or 'level of structured detail' .
      Also a measure of the quality of the regularized heights, because I have
      this sparse and dense point distribution...some heights are calculated with
      a dense neighbouring distribution, some with a sparse. As a pre-processing
      step I should somehow get a measure of 'expected interpolation accuracy' or
      something like that. The interpolation is done within a defined search
      window and using all available height in all octants with a weighted
      distance fuction. The a moving average plane will do the interpolation.
      That's just a short description of hwo it works.

      > error distribution obtained within a window. Finally, why not estimate the
      > variogram of the difference? This will tell you whether the error is
      autocorrelated.

      Unfortuately I don't have any experience in obtaining variograms and don't
      know if I can analyse such a large amount of data points (> 100000). If you
      have a hint on that, that would be useful !

      Thanks and best regards,
      Rainer Mueller
      --
      Dipl.-Ing. Rainer Mueller
      -------------------------
      German Aerospace Center - DLR Oberpfaffenhofen
      Remote Sensing Technology Institute
      Image Science - Photogrammetry Phone : +49 (8153) 28-2872
      PO Box 1116 FAX : +49 (8153) 28-1444
      82230 Wessling -GERMANY-
      EMail: Rainer.Mueller@...
      --




      --
      * To post a message to the list, send it to ai-geostats@...
      * As a general service to the users, please remember to post a summary of any useful responses to your questions.
      * To unsubscribe, send an email to majordomo@... with no subject and "unsubscribe ai-geostats" followed by "end" on the next line in the message body. DO NOT SEND Subscribe/Unsubscribe requests to the list
      * Support to the list is provided at http://www.ai-geostats.org
    Your message has been successfully submitted and would be delivered to recipients shortly.