terra

This R package is a replacement of the raster package. It has a very similar, but simpler, interface, and it is much faster.

All native computations are done in C++. Classes, methods and properties are exposed via a Rcpp module. The R side has two main S4 classes (SpatRaster and SpatVector) that represent spatial data. These classes have only slot, a reference to a C++ object. They are used to provide a "normal" "S4" R user-interface as in the raster package.

The first (alpha) release is expected by July 2019.

Installation

You need to install the latest version of "raster" from CRAN or github.

Windows

If you are on Windows, you need to first install Rtools to get a C++ compiler that R can use. After that, you need the first install development version of "raster" for "terra" to work.

Then, in R, install the packages.

library(devtools)
install.packages("raster")
#devtools::install_github("rspatial/raster")
devtools::install_github("rspatial/terra")

Mac - OSX

The libraries GDAL (>= 2.0.0), GEOS (>= 3.3.0) and Proj.4 (>= 4.8.0) are required (as for sf)

With Homebrew you can do:

brew install gdal

And now, in R, install the packages.

library(devtools)
install.packages("raster")
devtools::install_github("rspatial/terra")

Linux

The libraries GDAL (>= 2.0.0), GEOS (>= 3.3.0) and Proj.4 (>= 4.8.0) are required (as for sf)

To install these on Ubuntu you can do:

sudo add-apt-repository ppa:ubuntugis/ubuntugis-unstable
sudo apt-get update
sudo apt-get install libgdal-dev libgeos-dev libproj-dev 

And now, in R, install the packages.

library(devtools)
install.packages("raster")
devtools::install_github("rspatial/terra")

See the sf instructions for installation on other linux systems.