Mosaic creation of satellite data with FORCE in Python

The Python-based command line tool SADASADAM (SAme DAy SAtellite DAta Mosaics) makes it possible to create fully automated atmospherically corrected, cloud-free mosaics based on all available Sentinel-2 and Landsat-8/9 scenes on the same day.

Challenge

The automatic, standardized pre-processing of multispectral satellite data is a major challenge due to the enormous amount of data available and the different approaches to atmospheric correction and cloud detection depending on the sensor and provider. mundialis developed the Python-based command line tool SADASADAM to simplify the regular creation of Analysis-Ready-Data (ARD) mosaics of freely selectable study areas. This makes it possible to create atmospherically corrected, cloud-free mosaics of the same day based on all available Sentinel-2 and Landsat-8/9 scenes fully automatically for freely selectable study areas and time periods.
The Python package eodag is used internally to download the data, which serves as a standardized interface for various providers of satellite data (e.g. United States Geological Survey, Copernicus Dataspace). This avoids the dependence on individual providers and thus the potential instability of the tool, especially with regard to the shutdown of the Copernicus Open Access Hub.
Radiometric, topographic and atmospheric correction as well as cloud detection and mosaicking are carried out using FORCE software. FORCE breaks away from the concept of individual scenes and instead works with a data cube model, which simplifies the processing of large investigation areas and data volumes. In addition, processing can be optimized through parallelization depending on the available hardware.

Services

The company mundialis has fully implemented the project. The implementation took place in the following work steps:

  • Development of automatic Python download functions based on eodag
  • Development of a FORCE-Python interface
  • Development of Python post-processing functions based on GDAL
  • Development of a complete tool including command line interface
  • Creation of documentation on the use of the software
Result
  • Development of the Python-based command line tool “SADASADAM”
  • Implementation of software tests
  • Documentation and provision of the tool here
Customer
Fondazione Edmund Mach, San Michele All’adige (TN), Italy

Story

In the case of satellites that capture the earth’s surface in visible light and infrared, the amount of information is impaired by haze. These atmospheric effects can be reduced by appropriate data processing. The data comes in tiles and is compiled as a mosaic to cover the desired area. The computationally intensive processing could be optimized by FORCE. FORCE is an open-source software whose correction algorithms are recognized in the remote sensing community. As part of the SADASADAM (SAme DAy SAtellite DAta Mosaics) project, Mundialis developed a Python interface to FORCE to enable the FORCE modules to be called up efficiently as a command line tool for automated processing. The result: corrected, daily updated satellite data mosaics.
SADASADAM can be started using a single command, whereby the parameters are transferred via a configuration file. All FORCE parameters can be set by the user if required.

Atmosphere corrected Sentinel-2 mosaic of the study area Trento, Italy from 26/08/2023. Clouds have been removed.
Individual component of a SADASADAM result: Atmosphere-corrected, cloud-masked Sentinel-2 mosaic of the Trentino region. Many such individual time slices together form an analysis-ready data cube.
terrestris