Satellite images are the most comprehensive source of data about our planet. Data from space provides essential information on the Earth’s environment and the impacts caused by human actions. In recent years, space agencies have adopted open distribution policies; as a result, petabytes of Earth observation data are now available. Experts now have access to repeated acquisitions over the same areas; the resulting time series improve our understanding of ecological patterns and processes. Instead of selecting individual images from specific dates and comparing them, researchers and experts can track change continuously.
Satellite time series improve land use and land cover change analysis compared to what is achievable with an image from a single date or temporal composites. In addition, time series capture subtle changes in ecosystem health and condition and improve the distinction between different land classes. As a result, analysts can obtain the best benefits from big Earth observation data collections.
This book presents
sits, an open-source R package for land use and land cover classification of big Earth observation data using satellite image time series. The package provides an API to build EO data cubes from image collections available in cloud services and to classify these cubes using machine learning.
sits API provides an end-to-end toolkit for land mapping with Earth observation. Users can build regular data cubes from cloud services such as Amazon Web Services, Microsoft Planetary Computer, Brazil Data Cube, and Digital Earth Africa. The software includes quality assessment of training samples, machine learning and deep learning algorithms, and Bayesian post-processing methods for uncertainty assessment. Users can apply best practices for accuracy assessment of classification results.
The target audience for
sits is the community of remote sensing experts with Earth Sciences background who want to use state-of-the-art data analysis methods with minimal investment in programming skills. The package provides a clear and direct set of functions, which are easy to learn and master. Users with a minimal background on R programming can start using
sits right away. Those not yet familiar with R need only to learn introductory concepts.
If you are not an R user and would like to quickly master what is needed to run
sits, please read Parts 1 and 2 of Garrett Golemund’s book, “Hands-On Programming with R”(https://rstudio-education.github.io/hopr/>). If you already are an R user and would like to update your skills with the latest trends, please read the book by Hadley Wickham and Gareth Golemund, “R for Data Science”. (https://r4ds.had.co.nz/).
The version of the
sits package described in this book is version 1.2.0.
If you use sits in your work, please cite the following paper:
Rolf Simoes, Gilberto Camara, Gilberto Queiroz, Felipe Souza, Pedro R. Andrade, Lorena Santos, Alexandre Carvalho, and Karine Ferreira. “Satellite Image Time Series Analysis for Big Earth Observation Data”. Remote Sensing, 13, p. 2428, 2021. https://doi.org/10.3390/rs13132428.
This book is licensed as Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) by Creative Commons,, as described in the terms available https://creativecommons.org/licenses/by-nc-sa/4.0/. The
sits package is licensed under the GNU General Public License, version 3.0.