Sign inSign up

axensten/slu

By axensten

•Updated 2 months ago

PDAL filter modules to calculate z-based metrics, such as those used in forestry estimations.

Image
API management
Developer tools
Data science
0

10K+

axensten/slu repository overview

⁠Pax2 Docker Image

At the Forest Remote Sensing section of the Swedish University of Agricultural Sciences (SLU) we have developed a set of tools that we use to produce wall-to-wall rasters of Sweden for a set of forest variables. The tools are command line tools as well as filter modules for the command line tool PDAL⁠. The tools are packaged in this Docker image in a ready-to-run environment that also contains pdal, gdal, etc.

  • For more information on the pax2 tools and how you can use them, see the project’s githup page⁠. More documentation, the code, and a more detailed version history is available here.
  • For more information on how to use the pax2 tools to process files en-masse, see here⁠.

I hope you find the image useful.

Peder Axensten, Nyåker 2024

⁠Recent version history

  • 2024-05-18 Removed tree species from modelling. Removed previously tested models from code.
  • 2024-05-01 Now based on Ubuntu 24-04 LTS. Various small fixes.
  • 2024-03-05 New stable release. Various fixes.
  • 2024-02-09 Small fixes. It all seems to work now.
  • 2024-02-06 A reorganisation of the code. Needs a bit of internal testing before general use. Taged as "stable" by misstake -- sorry!
  • 2023-03-02 Now handles Biomass_above as well as Biomass_below. Still a test version using species information as stated in the 2023-02-24, above.
  • 2023-10-23 A bit of a rewrite, now entirely based on pdal in terms of point cloud handling. There are three main pdal modules: for raster metrics, for plot metrics, and for exporting points of plots. This version should be more flexible and faster. Upped the internal version number to 3.0.0.
  • 2023-07-25 Significant update. All point cloud related processing now use PDAL. Supplied in this Docker image are PDAL plug-ins for producing plot metrics, raster metrics, individual point clouds per each plot, to remove overlap due to different flight, and a special filter that we use at SLU to process point clouds from Lantmäteriet. The plot metrics calculation has also been opimized, so it should be considerable quicker. I plan to rewrite this page and focus it on the PDAL plug-ins, the make script to process large amounts of data is really secondary. I also plan to make tho code public on git-hub, but I need to remove unnecessary library dependencies first.
  • 2023-05-11 A new generalised system to define and calculate metrics. Now calculates bot Biomas_above and Biomass_below.
  • 2023-02-24 NOTE: experimental! Just to test it we are using species information in all models. It should not make estimations worse, but the question is if it improves estimations? Maybe for some variables.
  • 2023-01-25 Added a makefile variant 'extra-allt' that does what 'basic-linear' does, but also output a lot of extra metrics. If you only want to estimate variables as quickly as possible, use 'basic-linear'. If you want a lot of metrics to experiment with, use 'extra-allt'.
  • 2023-09-22 Don't use, for internal testing purposes only.
  • 2023-09-23 Implemented that a zero value 'max_distance' uses the radius from the plot table (column 'radius') and not the given value. The default value of it is zero.
  • 2026-07-08 Now on ubuntu:26.04. Various smaller fixes.
  • 2026-08-23 Merged two previous filters into one: plot_stuff. It both do metrics and save files with points from the plots. Now all points within or on the bounding box are processed. Previously, in C tradition, the lower bounds were included but not the upper. With floating points – the probability of this being an issue is small. But lately the Swedish Land survey started producing such files, so now also points on the upper limits are included.

Tag summary

Content type

Image

Digest

sha256:a960fe42b…

Size

990.3 MB

Last updated

2 months ago

docker pull axensten/slu