Forum

Author Topic: Classify Ground Points Settings and DEM generation  (Read 1878 times)

saellen

  • Newbie
  • *
  • Posts: 2
    • View Profile
Classify Ground Points Settings and DEM generation
« on: February 15, 2022, 12:35:53 AM »
Hi all,

I'm working on classifying the dense cloud to create a DSM/DTM in hopes of getting a canopy height model of tundra plants in Northern Alaska. The vegetation is between 5-50cm and the area is quite flat, so my needs are very fine scale. I have traditional air photos (stereotriplets) that were taken at 1:3000 and am using those to generate the dense cloud. I have been using the following parameters when classifying ground points:

max angle: 15 degrees
max distance: .05 m
cell size: .1 m

I then create a DTM from the classified ground points. Then I make a copy of the chunk and create a DSM from the UNCLASSIFIED points (excluding noise). Then I take it into arc and subtract the DTM from the DSM to get my CHM. The results seem totally reasonable minus a couple of negative values which I'm not sure why those are occurring. Some noise error I think?

I have another site where the vegetation is slightly taller, up to 90 cm. I am wondering if I should change the aforementioned parameters to something different when classifying the ground points. This site tends to have more shrubs (willows and dwarf birch) and less tussocks/graminoids.

So I am wondering:

1) does this general workflow seem appropriate for generating canopy height estimates in a tundra environment?
2) should I change the classify ground point settings for the site that has slightly taller vegetation or is this already pretty fine-scale?

Thanks for the input!!

Alexey Pasumansky

  • Agisoft Technical Support
  • Hero Member
  • *****
  • Posts: 14813
    • View Profile
Re: Classify Ground Points Settings and DEM generation
« Reply #1 on: February 15, 2022, 09:35:54 PM »
Hello saellen,

I think that 0.1 m for the cell size is too low value, as it is assumed that at least one ground point should exist in every cell of the initial grid.

But probably you can provide some screenshots of your results, like classified point cloud view and the results of DSM/DTM generation and their difference.
Best regards,
Alexey Pasumansky,
Agisoft LLC