Odeon-landcover’s documentation

ODEON stands for Object Delineation on Earth Observations with Neural network. It is a set of command-line tools performing semantic segmentation on remote sensing images (aerial and/or satellite) with as many layers as you wish.

Installation

These instruction assume that you already have conda installed.

First, download and extract a copy of odeon from repository Then navigate to the root of the odeon directory in a terminal and run the following:

Install and activate the environment

$ clone repository https://github.com/IGNF/odeon-landcover.git
$ cd odeon-landcover
$ conda env create --file=environment.yml
$ conda activate odeon
$ pip install .

In order to use cuda and NVIDIA devices, cudatoolkit must be installed too.

Quickstart

Odeon toolkit is run through main command:

$ odeon
  usage: odeon [-h] -c CONFIG [-v] {sample_grid,trainer}
  odeon: error: the following arguments are required: tool, -c/--config

Each tool needs a specific JSON configuration file. Available schemas can be found in odeon/scripts/json_defaults folder.

More information could be found in the documentation

Advanced Feature

API Reference

Indices and tables