Rapid image labeling for data-driven Earth science discovery.

A tool for web-based image annotation and efficient labeling pixels in images

Implements a rapid technique, described by Buscombe & Ritchie, (2018), for dense image labeling based on limited manual annotations.

Credits: Thanks to code contributions from Colin Talbert and Rich Signell

Example instructional video here

First, if you're a regular conda user, you should start by doing some housekeeping:

conda clean --all
conda update anaconda
conda update conda
conda update --all

Clone the repository:

git clone https://github.com/dbuscombe-usgs/EarthAnnotator.git

conda config --remove channels conda-forge

Create a conda environment:

conda config --add channels conda-forge
conda env create -f binder\environment.yml

Activate the conda environment (called EA):

conda activate EA

conda config --remove channels conda-forge

Add the kernel to jupyter:

python -m ipykernel install --user --name EA --display-name "Python (earthannotator)"

Have a look at the kernels

jupyter kernelspec list

Run the notebook (shows up in your web browser):

jupyter notebook