sits: Data Analysis and Machine Learning on Earth Observation Data Cubes with Satellite Image Time Series
2022-06-27
Preface
Using time series derived from big Earth Observation data sets is one of the leading research trends in Land Use Science and Remote Sensing. One of the more promising uses of satellite time series is its application to classify land use and land cover. Information on land is critical for sustainable development because our growing demand for natural resources is causing significant environmental impacts. As stated by [1], “time series analysis is expanding the kinds of land surface change that can be monitored using remote sensing. More subtle changes in ecosystem health and condition and related to land use dynamics are being monitored. The result is a paradigm shift away from change detection, typically using two points in time, to monitoring, or an attempt to track change continuously in time”.
This book presents sits, an open-source R package for land use and land cover classification using big Earth observation data. Users can build regular data cubes from collections in AWS, Microsoft Planetary Computer, Brazil Data Cube, and Digital Earth Africa. The software includes functions for quality assessment of training samples using self-organized maps. It provides machine learning and deep learning algorithms for classification of big Earth observation data cubes. Post-processing methods includes Bayesian smoothing and uncertainty assessment. Users can apply best practices for estimating area and assessing accuracy of land change. Thus, sits is an end-to-end toolkit for land mapping with Earth observation.
Who this book is for
The target audience for sits is the new generation of specialists who understand the principles of remote sensing and can write scripts in R. Ideally, users should have basic knowledge of data science methods using R. If you want more information on methods used in this book, please look at the following references:
Garrett Golemund, “Hands-On Programming with R”. O’Reilly, 2014. https://rstudio-education.github.io/hopr/.
Hadley Wickham and Gareth Golemund, “R for Data Science”. O’Reilly, 2017. https://r4ds.had.co.nz/.
Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani, “An Introduction to Statistical Learning with Applications in R”, Springer, 2013. https://www.statlearning.com/.
Zhang, A.; Zachary C. Lipton, Mu Li, Alexander J. Smola, “Dive into Deep Learning”, 2021. https://d2l.ai/.
How to use this book
This book describes sits version 1.0.0. Download and install the package as explained in the Setup section. Start at Chapter 1 to get an overview of the package. Then feel free to browse the chapter for more information on topics you are interested in.
| Chapter | Description |
|---|---|
| Chr 1 | Provides an overview of sits. |
| Chr 2 | Describes how to work with Earth observation data cubes. |
| Chr 3 | Describes how to access information from time series. |
| Chr 4 | Improving the quality of the samples used in training models |
| Chr 5 | Presents the machine learning techniques available in sits. |
| Chr 6 | Describes how to classify satellite images associated with Earth observation data cubes and smoothing methods to reclassify the pixels based on the machine learning probabilities. |
| Chr 7 | Presents the validation and accuracy measures. |
| Chr 8 | Presents case studies of LUCC classification. |
| Chr 9 | How to develop extensions to sits. |
Main reference for sits
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.