How to reproduce

1. install micromamba

For Linux, macOS, or Git Bash on Windows, install with: "${SHELL}" <(curl -L micro.mamba.pm/install.sh)

Default settings during the installation was fine for me.

2. create environment

micromamba create -n Chen2024

micromamba install -n Chen2024 -c conda-forge r-keras r-terra r-rsample tensorflow-gpu "keras<3" cuda=12.2 cudnn r-devtools r-sf r-lidr r-ggrepel

3. install NeonTreeEvaluation in R

micromamba activate Chen2024

R

devtools::install_github("Weecology/NeonTreeEvaluation_package", ref="c4d9953")

3 # do not update any package

quit()

n

micromamba deactivate

4. put function.R file in the right place

put function.R under ~/Chen2024/IPO/pro, which is the place it should be in for compatibility with the scripts. Otherwise, change paths accordingly in the scripts.

5. run the scripts with the created environment

Dataflow

The following figure provides an overview of the main folder structure, functions, and flow of data in this study.

Licence

This repository is licensed under the GNU General Public License, Version 3, or any later version.

In function.R, the function unet contained modification and the function conv2d_block contained copy from r-tensorflow/unet under the MIT License.

In function.R, the function unet contained modification from keras-team/keras under the Apache License, Version 2.0.