Phenotype Extraction

This R notebook is written to extract phenotype information from drone imagery using FieldImageR. This repo contains the following:

• Shapefile: shapefile\ containing the spatial polygons representing field plots.
• Fieldmap: fieldmaps\ containing the spatial organization of field plots.
• R notebooks: install_fieldimager.R, create_fieldmap.Rmd, create_shapefile.Rmd, extract_all_phenotypes_fieldimager.R

We will use these files to test FieldImageR

Table of Contents

• Download the GitHub repo
• Install FieldImageR
• Create fieldmap and shapefile
• Run phenotype extraction script

Download the GitHub repo

Download the repo by clicking Code > Download ZIP:

Now, extract the contents of the ZIP file:

Copy the path of the folder:

In RStudio, set the working directory to this copied path:

Install FieldImageR

To install FieldImageR, FieldImageR.Extra, and dependencies, run the install_fieldimager.R script.

Note: This script will also download red-green-blue (RGB) and digital surface model (DSM) orthomosaics from our 2024 maize trial.

Create fieldmap and shapefile

Before extraction, we must create a fieldmap and shapefile.

To create a fieldmap, run the create_fieldmap.Rmd notebook.

To create a shapefile, run the create_shapefile.Rmd notebook.

Note: Generating the fieldmap and shapefile is the responsibility of the graduate student or postdoc in charge of the project. These files are generated once per season and shared with the lab member who will be running processing.

Run phenotype extraction script

To run phenotype extraction, including vegetation indices, canopy cover, and plant height, run the extract_all_phenotypes_fieldimager.R file.

Note: Canopy cover and plant height are in metric units, specifically meters due to the input imagery being in the Universal Transverse Mercator (UTM) projection system.

As orthomosaics become available throughout the seasons, you can add them to 2024_maize_orthos. Make sure to place RGB orthomosaics in 2024_maize_orthos/rgb and DSM orthomosaics in 2024_maize_orthos/dsm. Once both RGB and DSM are added for a new flight, you can extract phenotypes from them using the extract_all_phenotypes_fieldimager.R file.

Note: You can edit the extract_all_phenotypes_fieldimager.R file to take inputs from another directory that is not maize specific. For example, 2024_cotton_orthos. However, keep in mind that the same subdirectories, 2024_cotton_orthos/rgb and 2024_cotton_orthos/dsm, need to be created and populated with data.