Segmentation_EB (Original Files)

new segmentation routine used for floe size distribution (FSD) and accurate floe characteristics (Buckley et al., 2024)

Files include:

• Segmentation_EB.ipynb - example notebook with commented seg algorithm code
• input
  • cloud - files for cloud mask
  • tci - images for analysis
  • fci - files for alternative cloud mask (not available yet)
• output
  • sample output

Packaged version

• Refactored and optimized code
• Notebook using packaged code
• CLI tool fsdproc for image processing

Installation

Preparation

python -m venv .venv # create an enviroment for running the package
source .venv/bin/activate # activate the enviroment just created
pip install --upgrade pip # upgrade pip in case it's an old/unsupported version

For regular use

pip install .

For development

Besides the standard package, extra tooling (e.g., testing, formatting, linting, coverage) can be installed with

pip install -e ".[dev]"

You can also use the included devcontainer, which preinstalls dependencies including:

• GDAL (for rasterio) and
• libgtk-3-dev (for cv2) Use the "Reopen in container" command in VSCode or as a GitHub codespace.

CLI

Upon installation the fsdproc command will be available. View its help with fsdproc --help.

Loading and Running

mkdir -p data/
fsdproc load data/tci.tiff --kind truecolor
fsdproc load data/cld.tiff --kind cloud 
fsdproc load data/lnd.tiff --kind landmask
fsdproc process data/tci.tiff data/cld.tiff data/lnd.tiff data/

To get data from Aqua, rather than terra:

fsdproc load data/tci.tiff --kind truecolor --satellite aqua
fsdproc load data/cld.tiff --kind cloud --satellite aqua