Terrarium

Terrarium is a Python Package that contains the Earth Engine and GIS related tooling for the GeoSentry 🌍 Platform.

Version: 0.4.1
Language: Python 3.9
License: MIT
Status: In Development

Overview

GeoSentry is a geospatial observation platform with spectral, topographical and comparative analytics along with a community-curated gallery of stunning satellite imagery powered by ESA's Sentinel-2 & Google Earth Engine.

Google Earth Engine

Terrarium uses the Earth Engine Python API for much of its geospatial manipulation functionality and satellite imagery datasources. Google Earth Engine is a planetary-scale platform for Earth science data & analysis that is powered by Google's Cloud Infrastructure and integrates well with Google Cloud Platform.

Satellite Sources

Terrarium is mostly confined to using the European Space Agency's Sentinel-2 MSI Level-2A dataset which is corrected for Surface Reflectance and available within the Earth Engine Data Catalog with the Collection ID COPERNICUS/S2_SR. The multi-spectral bands of the dataset are used for spectral index generation and false color compositions.

The Truecolor bands available to the L2A collection is also used for generation images, while the SCL band is used for rendering the Scene Classification Layer as part of GeoSentry's topographical analysis.

Cloudiness values are generated using Sentinel-2 Cloud Probability dataset which is also available on the Earth Engine Catalog with the Collection ID COPERNICUS/S2_CLOUD_PROBABILITY.

Terrarium also uses JAXA's ALOS DSM dataset for altitude based topographical analysis. The dataset is available with the Collection ID JAXA/ALOS/AW3D30/V3_2.

Installation

The package can be installed with pip using the following command.

pip install git+https://github.com/geosentry/terrarium#egg=terrarium

A specific version of the package can also be installed. The v0.3.0 tag can be installed using

pip install git+https://github.com/geosentry/terrarium.git@v0.3.0#egg=terrarium

Authentication

The package requires authentication setup to use services such as Google Earth Engine and Google Maps Platform.

Google Earth Engine

Authentication for the Google Earth Engine API is done using a IAM Service Account.

1. A Service Account needs to be created and registered with Earth Engin. Refer to the guide on creating and registering service accounts for Earth Engine.
2. Grant the Storage Object Admin IAM role to the Service Account. This allows it to perform exports to a Cloud Storage Bucket.
3. Generate a key for the Service Account and store it somewhere safe.
4. Set the path to this key file in the GOOGLE_APPLICATION_CREDENTIALS environment variable.
5. Terrarium uses this environment variable to point to the credentials used to authenticate the Earth Engine Session.

Google Maps Platform

Authentication for the Google Maps APIs are done using an API Key. The Terrarium package currently uses the Geocoding API for reverse coding and address resolution. API Keys can be created with the steps specified in this guide.

1. Create an API Key from the Google Cloud Platform Project
2. Store the API Key in the environment variable MAPS_APIKEY.

Changelog

The package changelog is located in the CHANGELOG.md file in the root directory of the repository.