Retrieve and process satellite images time series towards data analysis in R

Geomundus-2019, November 30, Manuel Montesino and Mehdi Moradi

Abstract

Satellite images are valuable sources to monitor the changes happening in the earth's surface patterns as well as dynamics. However, the analyses of such images often require a long series of remotely-sensed data. In two hands-on sessions, we first cover the use of the R package RGISTools to facilitate the handling of satellite images time series from major satellite programs, such as Landsat, MODIS, and Sentinel, covering in a comprehensive manner the retrieval, and customizing and processing satellite images. Having retrieved and processed satellite images time series of e.g. land surface temperature (LST) and/or vegetation indices from MODIS, we then review some of the most considered change-point and trend detection methods, and make use of them to discuss the areas where have faced changes over time together with the corresponding time of change for some provinces/cities in Europe.

Workshop program

15:15 - 16:15: Introductory walk-through to RGISTools for retrieving and customizing satellite imagery + exercises (Manuel); materials

16:15 - 17:15: Introduction to trend and change-point detection through different showcases, power of test and type I error, satellite images time series analysis + exercises (Mehdi); materials

Notes to keep in mind

1. Bring your laptop

2. Install the latest versions of R and RStudio

3. Install Google Earth

4. Packages used in part 1: devtools, RGISTools

    Install **RGISTools** as follows
   
    library(devtools)
    install_github("spatialstatisticsupna/RGISTools")
   

5. Packages used in part 2: trend, ecp, remote, gimms, repmis, plotKML, sp, zoo, raster