Hello - how do I go about filling raster gaps in a DSM?  I attempted to follow the approach used in Tutorial 8 (Working with Point Clouds) but, as indicated in tutorial, the approach is only good for filling small gaps and wasn't suitable for my application.   Thanks

4 comments

  • Nils Erik Jørgensen

    Hi James
    I have the same problem. I solved it by setting both dsm and dtm to the same value. I find it happened over lakes and lakes had no interest to me, so I did not care. Still I hope Trimble will fix this problem. It would be nice to have a good visual model for the whole area.

  • Matthias Stängel

    Hi James,
    Can you maybe share a screenshot? As far as I remember, a local approach is used, which makes sense. If you have large gaps such as lakes, you could classify those as such and simply assign one elevation value to the whole object instead of gradually interpolating from the edge of the object.
    Cheers

  • Christian Weise

    Hi James and Nils,

    Please take a look into the "fill gaps in elevation layers" rules of this example project:

    https://support.ecognition.com/hc/en-us/articles/360016224679-Tree-Crown-Delineation

    There many other approaches possible, but I hope this is a good starting point. The trick is to use the "pixel filter 2D" algorithm for the larger gaps or fill the lager gaps with a "fill pixel values" (linear interpolation = fast).

    Cheers
    Christian

  • Jarlath O'Neil-Dunne

    You may find it advantageous to generate the surface model outside of eCognition. eCognition has many great tools for loading in data. You can use a customized import to batch load data or the create/modify project algorithm to dynamically load the data as part of the rule set.

Recent Discussions

Need more help?

Enter Knowledge Base Contact Support