TreeSegmentation

An R package to replicate the submitted work using tree segmentation methods in R. This is a companion R package to the python-based https://github.com/weecology/DeepLidar.

https://www.biorxiv.org/content/10.1101/532952v1

This package follows the "Research Compendium" philosophy advocated by ROpenSci. It is not meant as a stand alone package, but rather a clear way to document our work.

Installation

Source can be cloned directly

git clone https://github.com/weecology/TreeSegmentation.git

Or installed directly into R.

library(devtools)
install_github("Weecology/TreeSegmentation")

Package structure

• \analysis - The main analysis scripts to generate trees from lidR clouds. These have relative paths which need to be amended by future users.
• \data - The publically available data. This is maintained for legacy purposed. We continue to update NEON evaluated trees https://github.com/weecology/NeonTreeEvaluation

Example

The main role of the package is to generate training bounding boxes for the deep learning analysis performed in the python.

See /analysis/detection_training