Filter points from LiDAR upon import to terrain model

I have LiDAR of an irregular stone object that I'm attempting to bring into PowerGEOPAK as a terrain, using the "Create from Point Cloud" tool.  It's working well, but my data needs cleaning up.  In the process of importing, it is filtering out 90% of the original data, which seems fine.  But I have many tiny spikes in the resulting surface.  There are too many to clean up by hand (I've tried).  Are there any parameters I can use during the import process that might smooth the surface, or at least remove isolated spikes?  I'm also working with Leica to see if there is any way to clean it up in Cyclone before attempting this with GEOPAK.  

Any advice would be helpful!  

Thanks,

Frank Boston, Michigan DOT

Parents
  • Hi Frank,

    I question if you've really reduced the point cloud by 90% using standard point cloud tools. I recommend getting Bentley Points Tools loaded on your machine. This software truly reduces the point cloud and you can easily edit the file after reduction. If your point clouds data is classified you can filter it using those classifications too. I struggled with Point Clouds until I started using this software.

    Idaho Red

  • Hi Red -

    First, the 90% reduction is precisely what Power GEOPAK reported upon import (from approximately 4M points to approx. 400K. 

    Second, the point cloud is not classified; it's a terrestrial scan. 

    My problem is really a smoothing issue; I need to find an automated way to remove the anomalous data points either before importing to a surface, or with surface editing tools after the import.  My problem is that all of the tools I'm used to using for flat paved surfaces don't work on such an irregular surface.  

    I've used Bentley's Point Tools in the past; I don't recall it having that kind of automated smoothing or editing, but I'll check it out again.  

    Thanks for the tip!

    Frank

Reply
  • Hi Red -

    First, the 90% reduction is precisely what Power GEOPAK reported upon import (from approximately 4M points to approx. 400K. 

    Second, the point cloud is not classified; it's a terrestrial scan. 

    My problem is really a smoothing issue; I need to find an automated way to remove the anomalous data points either before importing to a surface, or with surface editing tools after the import.  My problem is that all of the tools I'm used to using for flat paved surfaces don't work on such an irregular surface.  

    I've used Bentley's Point Tools in the past; I don't recall it having that kind of automated smoothing or editing, but I'll check it out again.  

    Thanks for the tip!

    Frank

Children