[pdal] Conifer Identification/ return points that are below certain height below max veg height

[Howard Butler]

> On Feb 29, 2024, at 6:37 AM, Sion Williams via pdal <pdal at lists.osgeo.org> wrote:
> 
> 
> I'm trying to find a sneaky/elegant way of distinguishing Broadleaf and Conifer trees from a 'Leaf Off' aerial survey made up of 27000laz tiles. Looking at a transect example of the veg class there are no lidar returns below a certain height for Conifers.(that are surrounded by other trees)
> 
> So any ideas how to filter(or add a dimension) the lidar below say 1.5m below the max returns(this could be referenced to a tif file for max height) and also the opposite for returning the top 1.5m values to then add to say a grouping filter. which could then go towards a tree count or a learning model.

I think you would have to resort to Python to compute these things at the moment, but PDAL 2.7.0 is going to add math operations, which would be very useful for doing something like I did in this gist https://gist.github.com/hobu/36ea5fbec36412965ad9ef44cf4ade48 

>     {
>         "type": "filters.assign",
>         "value": [
>             "MaybeTreeTops = 1"
>         ],
>         "where": "ReturnNumber == NumberOfReturns && NumberOfReturns > 1 && HeightAboveGround > 1.5"
>     },


This won't work with PDAL 2.6, but the idea is to write an indicator dimension and then use filters.expression to filter based on MaybeTreeTops==1. 


> 
> For conifer identification i have also looked at the lidar Intensity values but is a hit and miss process at the moment 

Hope this helps,

Howard

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/pdal/attachments/20240229/d8bb1115/attachment.htm>