LeafArea_calc

Measure leaf area from scanned photos with R and ImageJ, using the LeafArea package.

This is a GitHub template for the Plant Functional Trait Course. Do not work directly in the course template; make your own copy first.

Get a copy

Preferred: make a GitHub account, then use the template.

1. If you do not yet have a GitHub account, create one at github.com/signup. You need to be logged in for the Use this template button to appear.
2. On LeafArea_calc, click Use this template → Create a new repository.
3. Clone your new repository, or in GitHub click Code → Download ZIP and unzip it.

Without a GitHub account: you can still get the files. On LeafArea_calc, click Code → Download ZIP, then unzip the folder. You will not get a personal GitHub copy this way, but the R scripts work the same.

Unzip or clone the project into a path without spaces (for example C:/LeafArea_calc on Windows).

The workflow is the same on Mac and Windows. What differs is how ImageJ and Java are installed. Run the setup script once on each computer, then use the same analysis script everywhere.

Requirements

• R and RStudio
• Leaf photos in data/ (jpeg, jpg, tif, or tiff)
• No spaces in file or folder names (LeafArea cannot handle them)

You do not need Rtools, a separate Java install, or a manual ImageJ install. The setup script downloads original ImageJ (not Fiji / ImageJ2) with Java bundled.

Setup (once per computer)

1. Open LeafArea_calc.Rproj in RStudio so the working directory is this project.
2. Run code/00_setup.R.

That script:

• installs tidyverse and plyr from CRAN
• installs LeafArea from GitHub (the fork with extra crop options). This is a pure R package, so it is installed from a zip and does not need Rtools
• downloads the correct ImageJ build for Mac (Apple Silicon or Intel) or Windows
• unpacks it into tools/
• checks that ImageJ's Java actually runs

On Mac, ImageJ lives at tools/ImageJ.app. On Windows, it lives at tools/ImageJ.

Calculate leaf area

1. Put images in data/ (subfolders are fine).
2. Run code/01_calculate_leaf_area.R.
3. Check output/leaf_area.csv and the masks in output/masks/.

Each image is processed one at a time. If one scan fails, the rest still run.

output/leaf_area.csv has one row per scan:

• id: file name without extension (e.g. DKQ6258)
• n_particles: number of objects ImageJ counted as leaves
• leaf_area: total area in cm²

If n_particles is much larger than the number of leaves, ImageJ probably picked up the envelope, ruler, tape, or dirt. Inspect that scan's mask and adjust the trim settings (below).

Image requirements

Photos should be 300 dpi. The default scale is 237 pixels = 2 cm, which matches 300 dpi scans used in previous trait workflows.

These photos include an envelope at the top and a ruler on the right. The script crops those away before measuring. The leaf should sit on a light background.

Settings you may need to change

In code/01_calculate_leaf_area.R:

Setting	Default	Meaning
distance_pixel / known_distance_cm	237 pixels = 2 cm	Scale. Change only if scans are not 300 dpi.
trim_pixel	58	Crop a little from the edges.
trim_pixel_right	150	Crop the ruler.
trim_pixel_top	1500	Crop the envelope.
low_size	0.1	Ignore specks smaller than this (cm²).

If the mask still includes the envelope, increase trim_pixel_top. If it includes the ruler, increase trim_pixel_right. If a small leaf is missing, lower low_size.

Mac vs Windows

code/01_calculate_leaf_area.R is the same on both systems. Differences are only in setup:

• Mac: ImageJ is ImageJ.app. LeafArea calls java on the PATH, so the scripts point JAVA_HOME at ImageJ's bundled JRE. This avoids the macOS /usr/bin/java stub, which is not a real Java runtime.
• Windows: ImageJ is a folder containing ij.jar and jre\bin\java.exe. LeafArea uses that java.exe directly.

Do not install Fiji. LeafArea only supports original ImageJ.

Folders

LeafArea_calc.Rproj
code/
  00_setup.R                 # run once: packages + ImageJ
  01_calculate_leaf_area.R   # measure leaf area
  imagej_helpers.R           # shared Mac/Windows ImageJ helpers
data/                        # put leaf images here
output/
  leaf_area.csv
  leaf_area_raw.csv
  masks/                     # ImageJ outlines for checking
tools/                       # ImageJ is downloaded here (not committed)

Troubleshooting

Rtools is required / Rtools is not available for this version
Rtools is only needed to compile packages (C/C++/Fortran). LeafArea is pure R, so this project does not need it.

remotes::install_github() still checks for Rtools on Windows and can fail even when compilation is unnecessary. Setup therefore installs the GitHub zip with install.packages(), which does not require Rtools.

If you still want Rtools for other packages: R 4.4 needs Rtools44, R 4.5 and R 4.6 need Rtools45 in C:\rtools45. There is no Rtools46.

The scan name prints, but there is no leaf area (Windows)
On Windows, ImageJ is started from the ImageJ folder, so it cannot see a relative folder like data_temp. The script now passes an absolute path. Re-run the whole of 01_calculate_leaf_area.R from the top. If a black Command Prompt window says Press any key to continue, press a key; that pause comes from LeafArea.

Also avoid spaces in the project path (C:/LeafArea_calc is safer than C:/Users/First Last/Documents/...).

Unable to locate a Java Runtime
Run code/00_setup.R, then re-run the whole of 01_calculate_leaf_area.R from the top (not just the last few lines).

ImageJ not found
Run setup again, or download ImageJ from https://imagej.net/ij/download.html (the build bundled with Java) and place ImageJ.app (Mac) or the ImageJ folder (Windows) in tools/.

Area looks wrong / too many particles
Open the matching file in output/masks/ and compare it to the original photo. Adjust trim or low_size, then re-run.

Setup download fails
Download the zip yourself from https://imagej.net/ij/download.html, unzip it into tools/, and run 00_setup.R again. It will reuse the local copy.