VegMapper

Land cover classification using remote sensing observations.

Prerequisites

If you own a cloud environment, follow steps 1-5 below. If you are using OpenSARlab, go straight to step 4.

The VegMapper software is intended to be used on a cloud-computing platform (e.g., AWS EC2) as the volume of remote sensing data is very large normally. However, VegMapper can still be used on a local machine but only Linux and macOS platforms have been tested.

1. Obtain an AWS account and EC2 instance

Follow steps on CLOUD.md.

2. Install Git and Conda

• Git (for cloning this repository to the machine used)

• Miniconda (for installing other required Python packages)

After Miniconda is istalled, install jupyterlab, mamba, and kernda into the base environment by running:

conda install -n base -c conda-forge jupyterlab mamba kernda

3. Running JupyterLab on a remote server (EC2)

• From local machine, run the following command:

  ssh -L 8080:localhost:8080 <EC2_username>@<EC2_address>
  

  to connect to EC2 with the port forwarding.

• After logging into EC2, run the following command:

  jupyter lab --no-browser --port=8080
  

  to launch JupyterLab.

• After launching JupyterLab, in the output message it will provide URLs for accessing the Jupyter server. Copy and paste one of them (for example, http://localhost:8080/lab?token=your_token) in your browser to open the Jupyter Lab web interface.

4. Clone VegMapper Repository

Open a terminal, navigate to where you want the repository to be cloned to, and do

git clone https://github.com/NaiaraSPinto/VegMapper.git

5. Installation

To create a conda environment and install VegMapper software, navigate to the VegMapper folder and run this Jypyter notebook: INSTALL.ipynb. This notebook also includes instructions for obtaining credentials to download imagery from NASA and JAXA.