1

I have a LiDAR point cloud that has a density of 16.4 points per m^2 and spacing of 0.2469m, and am trying to extract trees from it but the classification came with a lot of messy points in classes they shouldn't be in, so I am trying to reclassify it from scratch using GlobalMapper V21.

My problem is that when I try and classify the buildings and vegetation I am getting building roof outlines in the high vegetation, though the main building roofs do show up in the buildings classification. I have tried many permutations of the classification filtering settings and have improved things a bit, but I still can not get these building outlines out of my vegetation layer.

I have read the GM help page on what the settings do but am still scratching my head as to the best values to use. The ground classification help page had good descriptions and example pictures of results using differing values, but the building/veg classification pages don't have the same degree of explanation so i am finding it hard to trial & error my way to a good result.

What suggestions do you have for improving this so I get buildings in the building class and only trees in the high & medium vegetation classes?

Attached is a picture of the best result I can get showing the building outlines, but also the settings seem to miss some of the trees.enter image description here

2
  • 'My problem is that when I try and classify the buildings and vegetation I am getting building roof outlines in the high vegetation, though the main building roofs do show up in the buildings classification.' It seems buildings are classified correctly, is that right? If so, split your point cloud (through class codes) in 'building points' and 'all other class, except buildings'. Classify the 'all other class, except buildings' point cloud from scratch. Then, merge it with the split 'build-only' point cloud. See if results are better. Apr 1, 2020 at 14:53
  • I would suggest contacting Blue Marble's tech support. They can most assuredly help you.
    – BlueMax
    Jun 5, 2020 at 15:17

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.