Skip to content

Home

A python Earth Observation toolbox

Conda Version Conda Downloads

EO-Tools is a pure python toolbox that is currently able to search, download and process Sentinel-1 InSAR pairs, geocode and apply terrain correction to Sentinel-1 SLC products, download and mosaic Sentinel-2 tiles and download various publicly available DEM (Digital Elevation Models). The S1 processor can compute phase, amplitude and coherence in the SAR geometry and reproject them in a geographic coordinate system. Example notebooks demonstrating the different features are located in the notebooks-cf folder of the github repository.

New

EO-Tools can now download and process Sentinel-1 SLC partial products. Instead of downloading an entire Sentinel-1 product, the partial-product downloader retrieves the metadata and only the consecutive bursts intersecting an area of interest. The result is stored as a SAFE-like <product-id>.partial.SAFE directory and can be passed directly to the existing Sentinel-1 processors.

Partial products provide several advantages:

  • Smaller downloads and reduced local storage requirements.
  • Faster access to the Sentinel-1 data needed for a specific area.
  • The download AOI, available subswaths, polarizations, and burst ranges are stored with the product for reproducible processing.
  • Existing high-level InSAR, SLC geocoding, H-Alpha, and polarimetric covariance processing APIs automatically recognize valid partial products.

See the partial-product download tutorial and partial-product processing tutorial to get started.

Examples

Here are examples of EO-Tools outputs showing amplitude, coherence and inteferometric phase of a 2023 earthquake in Morocco,

Amplitude. Coherence. Phase.

a comparison between Sentinel-1 amplitude, coherence, change map using two dates in 2021 and Sentinel-2 RGB image over the city of Berlin in Germany,

Amplitude Berlin. Coherence Berlin. Change Berlin. S2 Berlin.

and the comparison between Beta nought calibration and terrain flattening using Copernicus DEM on Sentinel-1 data over Etna.

Beta0. Flattened.

Overview

  • Currently, the available features are:
    • Sentinel-1
      • Interferometric processing of Sentinel-1 pairs, including TOPS processing steps like azimuth deramping, DEM assisted coregistration, Range-Doppler terrain correction and Enhanced Spectral Diversity. Individual bursts can be processed as well as full products and cropped to any area of interest provided by the user
      • Amplitude geocoding of SLC Sentinel-1 images, with Beta or Sigma Nought or terrain flattening calibration.
      • Ability to apply processing in the SAR geometry and further project the results in a geographic coordinate systems using lookup-tables.
      • Writing the result as a geocoded (terrain corrected) COG (Cloud Optimized GeoTIFF) file
      • Displaying these rasters on top of a folium map in a jupyter notebook (docker version only)
    • Sentinel-2
      • Tile merging and geocoding
      • Write any band to COG files
      • Visualization of color composites (Natural RGB, CIR, SWIR, etc) on a folium map (docker version only)
    • DEM
      • Automatically downloads and crops a DEM given a geometry
    • All products
      • Search catalog (using EODAG) and download products
      • Explore products by displaying their footprint on a folium map (custom function)
      • Show remote and local images on top of folium maps in the notebook
  • Example notebooks can be found in the notebooks/ folder

Install & quick start

  • The package comes in two flavors
    • A conda package that contains the main functionality (Sentinel-1 InSAR, Sentinel-2 tile mosaic and DEM download)
    • A docker version (for more advanced users) that additonally works with a TiTiler server for interactive visualization in the notebooks
  • It is recommended to first create a conda environment to avoid package conflicts
  • You need to have conda installed (or mamba / micromamba)
  • Then the package can be installed with these commands (replace conda by mamba or micromamba if needed):
conda create -n eo_tools
conda activate eo_tools
conda install conda-forge::eo-tools

Docker install

  • It works as a dev container for VSCode.
    • Clone the github repository into the location of your choice.
    • Volumes paths can (and should) be changed in docker-compose.yml.
    • After opening the main directory, VSCode should detect the devcontainer file and ask to build the container. Once the container is running, the example notebooks in the notebooks directory can be used.
  • Alternatively, it should also be possible to start the container from the main directory with docker-compose up -d in a terminal and attach to the container with any editor supporting docker.

Getting started

  • Please make sure jupyter is installed in your environment
  • Example jupyter notebooks demonstrate the different features
  • For conda use the notebooks in the notebooks-cf directory of the github repository
  • For docker use the notebooks in the notebooks directory of the github repository
  • To be able to download products with EODAG on Copernicus Dataspace, a file named eodag_config.yaml containing credentials must be created. The file eodag_config_example.yml located in the repository root directory may be used for this purpose. Feel free to copy, rename and edit this file with your credentials. It is recommended to store the file in a secure location outside of the local repository.

Documentation

Documentation is available on Read the Docs.

Notice

  • This project was originally forked from: https://github.com/eo2cube/s1_processor/, however since 99% of the code is now original, I have detached the fork.
  • Visualization functions are using TiTiler https://developmentseed.org/titiler/
  • Full-product discovery and download use EODAG https://github.com/CS-SI/eodag
  • Sentinel-1 partial-product discovery and download use the Copernicus Data Space Ecosystem STAC and S3 services

Some example notebooks for the conda version

Some examples can be viewed here: