1 - Installing ROpenCVLite

Simon Garnier

Installing ROpenCVLite should go smoothly on all major operating systems (Mac, Windows, and most major Linux distributions). There are a few steps to follow before installing the package itself in order to prepare all the necessary external (i.e. non-R) dependencies for OpenCV. You should also make sure that your R version is 4.0 or higher. Support for older versions of R has been removed.

All installation steps are described below.

1 - Before installing

1.1 - Windows

Before installing ROpenCVLite, you will need to install Rtools and CMake.

1.1.1 - Rtools

You will need to install the version of Rtools that corresponds to your version of R. You can check your version of R by typing R.version in the R console. The Rtools download page always lists the current recommended version for each R release.

There are multiple ways to install Rtools on Windows:

Option 1

Use the convenience function provided by the installr library:

if (!require("installr"))
  install.packages("installr")

installr::install.Rtools()
Option 2

If you have the rig utility installed on your computer, you can install Rtools from a terminal running PowerShell (replace rtools44 with the version matching your R installation):

rig add rtools44
Option 3

If you have the winget utility installed on your computer, you can install Rtools from a terminal running PowerShell:

winget install RProject.Rtools
Option 4

You can manually download Rtools at https://cran.r-project.org/bin/windows/Rtools/.

1.1.2 - Cmake

Installing Cmake is only required if your R version starts with 4.0 or 4.1 and, therefore, you have installed Rtools40. For newer versions of R, Cmake is included in Rtools42 and Rtools43 and you can skip this step.

Any version of Cmake 3.24 or higher should work. There are multiple ways to install Cmake on Windows:

Option 1

Use the convenience function provided by the installr library:

if (!require("installr"))
  install.packages("installr")
  
installr::install.CMake()

During the installation process, make sure to tell the installer to add CMake to your “PATH”.

Option 2

If you have the winget utility installed on your computer, you can install Cmake from a terminal running PowerShell:

winget install cmake
Option 3

You can manually download the CMake installer for Windows at https://cmake.org/download/. Make sure to tell the installer to add CMake to your “PATH” during the installation process.

1.2 - Mac

Before installing ROpenCVLite, you will need to install the latest version of CMake. There are multiple ways to install Cmake on Mac.

Apple Silicon (M-series Macs): ROpenCVLite is fully supported on arm64 hardware. The package will compile a native arm64 build of OpenCV automatically. If you are running an x86_64 version of R under Rosetta 2, the build will target x86_64 instead. To get a native arm64 build, use an arm64 build of R (available from CRAN as the “arm64” installer).

Option 1

If you have homebrew installed, run the following command in your terminal:

brew install cmake
Option 2

If you have MacPorts installed, run the following command in your terminal:

sudo port install cmake
Option 3

Go to https://cmake.org/download/. Scroll down until you reach the “Latest Release” section. Download the installer for Mac OS X (.dmg file). Once the download is complete, double-click on the installer and accept the license agreement. Drag and drop the Cmake app into your Applications folder.

To make Cmake accessible to R, open a terminal and type:

sudo "/Applications/CMake.app/Contents/bin/cmake-gui" --install

It will ask for your administrator password.

1.3 - Linux

Before installing ROpenCVLite, you will need to install the latest version of CMake. Hereafter are two options to do this on Linux.

Use your distribution’s package management system. For instance on Ubuntu-like systems, you can install Cmake by typing the following command in your terminal:

sudo apt install cmake

Other distributions will have different package managers. Refer to the documentation of your distribution for finding the proper way to install Cmake on your system.

Download CMake for Linux at https://cmake.org/download/. However it is recommended that you install it using your distribution’s package management system.

2 - Installing ROpenCVLite

You can install ROpenCVLite from CRAN as follows:

install.packages("ROpenCVLite")

Or you can install the latest development version (not recommended) as follows:

# Using pak (recommended)
pak::pak("swarm-lab/ROpenCVLite")

# Or using remotes
remotes::install_github("swarm-lab/ROpenCVLite")

This will not install OpenCV yet. The first time you load the package, it will notify you if OpenCV needs to be installed or updated:

library(ROpenCVLite)

To install or update OpenCV, run:

ROpenCVLite::installOpenCV()

This step can take some time. Go make yourself a cup of coffee or two while ROpenCVLite downloads, compiles and installs OpenCV on your system.

ROpenCVLite will also automatically detect when a new version of OpenCV is available and notify you at startup.