A PovRay-powered tree interpolator for creating forest digital surface models from LiDAR point clouds.
This is a project which puts verisimilitude at or above the level of veracity, i. e. it will create a surface model of a forest canopy down to the last little leaf, in the absence of that level of detail in the LiDAR dataset itself.
That said, the output data do (also) have a high level of veracity.
This output is accomplished by using (effectively) a 3D tree "stamp" which is placed at each and every LiDAR point, and scaled to the height above groun of that point. See https://smathermather.wordpress.com/tag/lidar/ for more info.
Requirements:
- Linux Machine Running Wine
- rapidlasso's LASTools: http://rapidlasso.com/lastools/
- PovRay 3.6 or later: http://povray.org/
- Bourne Again Shell (BASH) (make sure PovRay, LASTools, tree.inc, and treepov.inc are included in your path)
Functionality
Written (ATM) in BASH, this code does a few simple things:
- Uses lasheight to calculate the height of all points above ground
- Uses las2las to remove class 1 and 2 in the LiDAR data, leaving class 5 (tall vegetation)
- Converts the las to text and creates:
- File of coordinates of XY laspoints, as a PovRay array and an array of point heights from ground. XY array is used to place tree "stamp", point height array used to scale stamp size in XYZ dimensions
- PovRay File for rendering (.pov) to PNG (.png)
- Executes the render of the PNG
- World file (*.pgw) to reference the PNG into your coordinate system.
tree.inc is a triangular mesh of a tree generated using Pov-Tree, now defunct but available at the illustrious Internet Archive: http://web.archive.org/web/*/http://propro.ru/go/Wshop/povtree/povtree.html (in case you want a different tree shape). treepov.inc is the additional code needed create the renderable povray document
Usage
tree-interpolator.sh input.las output-name width-height pixel-sizee.g.
./tree-interpolator.sh input.las Licking 5000 10 Outputs
The primary outputs from this script are, as mentioned in functionality, are a PNG and a world file that georeferences that PNG. The data are scaled between 0 and 63575, and the code assumes the tallest trees are 175 feet tall. While this is a reasonable assumption for Ohio, it may not be reasonable for e.g. California.
