Struggling to isolate high vegetation in LIDAR data
JamesRutter
Global Mapper UserTrusted User
Any observations on how I can improve the classification of 'High Vegetation' in my LIDAR data?
I have accurate LIDAR at 0.5m post spacing. I've done an automatic classification of ground points and have moved on to classify non-ground points to see if I can extract all tree canopies. Having a few problems though with a lot of building stuff being picked up and classified as high vegetation. I'm not sure how to tweek the autoclassify settings to improve and clean up the classification process, or whether there are some more filtering settings which might improve the process (eg filtering on colour or scan angle (not sure what scan angle is though!)).
Screenshots attached show a flat view of my attempt filtered to just what GM thinks is 'high vegetation', a 3D view of the same info plus the autoclassification settings. Most of the stuff in the middle of the vertical view is building structures in the town centre which are getting picked up and classified as 'high vegetation'.
Thanks
JamesLidar1.jpgLidar2.jpgLidar3.jpg
I have accurate LIDAR at 0.5m post spacing. I've done an automatic classification of ground points and have moved on to classify non-ground points to see if I can extract all tree canopies. Having a few problems though with a lot of building stuff being picked up and classified as high vegetation. I'm not sure how to tweek the autoclassify settings to improve and clean up the classification process, or whether there are some more filtering settings which might improve the process (eg filtering on colour or scan angle (not sure what scan angle is though!)).
Screenshots attached show a flat view of my attempt filtered to just what GM thinks is 'high vegetation', a 3D view of the same info plus the autoclassification settings. Most of the stuff in the middle of the vertical view is building structures in the town centre which are getting picked up and classified as 'high vegetation'.
Thanks
JamesLidar1.jpgLidar2.jpgLidar3.jpg
Comments
-
Further queries. It seems that if I now colour up my filtered 'high vegetation' points (which still include all the spurious sides of buildings etc which I want to get rid of) using the 'Color LIDAR by Return Height Delta' most of my trees are displayed red and all the spurious building stuff which I want to filter out get displayed red or green. First off what is the 'return height delta' and is there any way in GM to filter the LIDAR on this (presume there must be a method if it can color the points using this attribute)? Secondly I was thinking that if this doesn't work I might be able to filter out the points by colour as I've coloured my LIDAR point cloud from an underlying RGB air survey. In the attached screenshot of the dialogue box showing LIDAR point attributes, what are teh red,green and blue values? they don't look like normal RGB values to me??Lidar4.jpg
Categories
- 12.7K All Categories
- 5.6K Features Discussion
- 342 Downloading Imagery
- 1.3K Elevation Data
- 380 Georeferencing Imagery Discussion
- 627 GM Script Language
- 52 User Scripts
- 113 GPS Features
- 414 Projection Questions
- 819 Raster Data
- 1.3K Vector Data
- 6.6K Support
- 177 Announcement and News
- 908 Bug Report
- 558 SDK
- 1.2K Suggestion Box
- 3.7K Technical Support
- 562 Other Discussion
- 129 GIS Data Sources
- 27 Global Mapper Showcase
- 233 How I use Global Mapper
- 107 Global Mapper Forum Website