DHI-GRAS/efast

Efficient Fusion Algorithm across Spatio-Temporal scales (EFAST)

EFAST: Fusion of Sentinel-2 and Sentinel-3 Data

Introduction

The EFAST package is designed for combining Sentinel-2 and Sentinel-3 data to produce frequent, high-resolution images. This approach can also be applied to other satellite datasets, such as Landsat and MODIS. The purpose of this package is to provide analysis-ready data, that is, cloud-free optical images at regular time-steps.

Motivation

Sentinel-3 satellites provide daily images of the same area with a coarse resolution of about 300 meters on the ground. On the other hand, Sentinel-2 images have a higher resolution of up to 10 meters but have a longer revisit time. By combining the two satellite datasets, it is possible to obtain time-series of both high temporal and spatial resolution. This makes the EFAST package a valuable tool for monitoring ecosystems and extracting key information.

Use cases

The EFAST package is intended for users who:

1. Monitor ecosystems using Sentinel-2 or Landsat but are limited by the long revisit time of these satellites.
2. Want to make full use of the synergy between Sentinel-2, Sentinel-3, Landsat, and MODIS to obtain high-resolution time-series images.
3. Need cloud-free optical images at regular time-steps for their analysis.

Reference

Senty, P., Guzinski, R., Grogan, K., Buitenwerf, R., Ardö, J., Eklundh, L., Koukos, A., Tagesson, T., and Munk, M. (2024). Fast Fusion of Sentinel-2 and Sentinel-3 Time Series over Rangelands. Remote Sensing 16, 1833. https://doi.org/10.3390/rs16111833.

NDVI Example: Aarhus, Denmark in Spring 2021

The EFAST package can be used to generate cloud-free NDVI (Normalized Difference Vegetation Index) images, as demonstrated by the example of Aarhus, Denmark in Spring 2021.

How to Use EFAST

See run_efast.py for an example using data located in test_data folder.

Requirements

• python
• esa-snap - needed for Sentinel-3 pre-processing only. Tested with version 9 and 10.

Try it out

1. Clone the repository to your local machine.
2. Navigate to the root directory of the repository in your terminal.
3. [OPTIONAL but recommended] Create a virtual environment: python3.<your python version> -m venv .venv
4. Install the package dependencies: pip install -r requirements.txt
5. Change the credentials in run_efast.py to your own CDSE credentials (needed to download Sentinel-2 and Sentinel-3 data).
6. Run the example: python run_efast.py

Installation

Install the package using pip:

pip install git+https://github.com/DHI-GRAS/efast.git

Usage

import efast

...
efast.fusion(
    ...
)

Develop

1. Clone the repository to your local machine.
2. Navigate to the root directory of the repository in your terminal.
3. [OPTIONAL but strongly recommended] Create a virtual environment: python3.<your python version> -m venv .venv
4. Install the package in dev mode: pip install -e .[dev]