I am looking for any source code that does tree crown segmentation/delaunitaitoun with LiDAR point clouds or their respective greo-TIFFs. So far I can load and view both LiDAR file and its TIFF in matlab but I need the code to convert it to extract tree crowns.

Any open source library in any programming language is ok.

  • 1
    What method do you want to apply? There are a number of papers concerning crown extraction from lidar data (wavelets, nonlinear regression, watershed segmentation, Gaussian decomposition, ect...). It would be helpful if you narrowed it down a bit. Commented Feb 2, 2015 at 16:29
  • I am working on providing scalable solutions for lidar data processing (on terrabytes scale). I am looking to take on a simple crown segmentation source code and develop a scalable version of it. Commented Feb 2, 2015 at 16:36

5 Answers 5


A literature search would provide you a wealth of information!

Bob McGaughey with USFW-PNW in Seattle, is the developer of FUSION and I am sure would hand over the source code for watershed segmentation.

Randy Wynne is at Virgina Tech and developed an IDL virtual machine program implementing a variable window filtering approach Popescu & Wynne (2004). IDL is a fairly transparent language and would be easy to translate to a different language. If you need to track him down, I believe that Popescu is now at Texas A&M.

Finally, Mike Falkowski at University of Minnesota published on using wavelets for tree crown extraction and has an implementation in Matlab that he may be willing to share.

Please keep in mind, it is all about application. Take stock of what system you are working in before settling on an algorithm. There are a few papers that compare methods and demonstrate that specific algorithms work better in certain types of systems. For instance, VWF does not work as well in dense, multi-storied canopy as in open forests.

  • Thanks for the comment. I check them out and update here asap. :) Commented Feb 2, 2015 at 17:03
  • Bob McGaughey was kind enough to share part of his code, which I'm grateful for. Thank you Commented Feb 3, 2015 at 17:27

You can check out these web-based applicatons:

  1. LiDARTreeTop http://forest.moscowfsl.wsu.edu:3838/LiDARTreeTop/

  2. LiDAR3DclusterTree http://forest.moscowfsl.wsu.edu:3838/LiDAR3DclusterTree/


To visualize, you can check out these web-based applications:

  1. LiDARtreesModel3D http://forest.moscowfsl.wsu.edu:3838/csilva/LiDARtreesModel3D/

  2. LiDARstand3D http://forest.moscowfsl.wsu.edu:3838/csilva/LiDARstand3D/

  • Since you already have "lollypop" trees built in LiDARstand3D are you deriving crown bulk density? If you calculate CBD, the fire folks would absolutely love you. Fairly short step from where you are. I always had very good correlations with coefficient of variation of vertical canopy density and crown competition factor, but you should be able to get directly at that as well. Nice work. Commented Feb 23, 2015 at 22:23
  • Cool stuff! Is the source code available to build 'academic' systems on top of them, where we would cite your work as building block? Commented Feb 24, 2015 at 20:51

Also, you could get the tree canopy volume using this web-based application:

  1. LiDARAlphaShape3D http://forest.moscowfsl.wsu.edu:3838/LiDARAlphaShape3D/

Also, check LAStools: http://www.cs.unc.edu/~isenburg/lastools/ for rapid and clean processing of lidar data. This link will show you how lastools can be used in a workflow for tree crown calculations: http://rapidlasso.com/2014/10/23/discriminating-vegetation-from-buildings/

  • 1
    There are no crown extraction algorithms implemented in LASTools. Commented Feb 2, 2015 at 18:43

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