ashutoshkumarjha / pygeo

pyeo

Python for Earth Observation

This is designed to provide a set of portable, extensible and modular Python scripts for machine learning in earth observation and GIS, including downloading, preprocessing, creation of base layers, classification and validation.

Full documentation at https://clcr.github.io/pyeo/build/html/index.html

Example notebooks are at https://github.com/clcr/pyeo_training_materials

Requirements

Package management is performed by Conda: https://docs.conda.io/en/latest/

For downloading, you will need a Scihub account: https://scihub.copernicus.eu/

For Sentinel 2 processing, you may need Sen2Cor installed: http://step.esa.int/main/third-party-plugins-2/sen2cor/

For AWS downloading, you will need credentials set up on your machine.

To install

To install Pyeo, put the following commands into Bash (Linux), Terminal (Mac) or the Anaconda Prompt (Windows)

git clone https://github.com/clcr/pyeo.git
cd pyeo
conda env create --file environment.yml --name pyeo_env
conda activate pyeo_env
python -m pip install -e .

If you want access to the Pyeo command line functions, add the following to your .bashrc

export PYEO=/path/to/pyeo
export PATH=$PATH:$PYEO/bin

If you do not want to edit Pyeo, replace the pip install line with

python -m pip install . -vv

You can test your installation with import pyeo.classifier

Example script

This presumes a set of training data exists, you have signed up to Scihub and the folders s2_l1, s2_l2, preprocessed and classified have been created.

from pyeo import raster_manipulation as ras
from pyeo import queries_and_downloads as dl
from pyeo import classification as cls

# train_model.py
cls.extract_features_to_csv("training_raster.tif",
                            "training_shape.shp",
                            "features.csv")
cls.create_model_from_signatures("features.csv", "model.pkl")


# classify_area.py
username = "scihub_user"
password = "scihub_pass"
conf = {'sent_2':{'user':username, 'pass':password}}
data = dl.check_for_s2_data_by_date("aoi.shp",
                                    "20200101",
                                    "20200201",
                                    "conf")
dl.download_s2_data(data, "s2_l1", "s2_l2", username, password)
ras.preprocess_sen2_images("s2_l2", "preprocessed", "s2_l1")
cls.classify_directory("preprocessed",
                       "model.pkl",
                       "classified",
                       apply_mask=True)

(This is a toy script; keeping your username and password in your script is not recommended in the real world).