Installation of Computer Vision tools

These are some of the most commonly used Computer Vision tools in Robotics :

virtualenvwrapper

Virtualenvwrapper is a set of extensions and wrappers used for creating, managing and deleting Python virtual environments using the terminal. Virtual environments are isolated Python installations that help in keeping dependencies required by different projects seperated. They also help in keeping a track of the dependencies or packages required by a project during deployement.

You can read more about it here(1)

OpenCV

OpenCV (Open Source Computer Vision Library) is an open source Computer Vision and Machine Learning software library, originally developed in C++. It is one of the most commonly used Computer Vision tool nowadays, with client libraries available in popular languages like Python, Java, MATLAB etc.

You can learn more about it here(2)

cv_bridge

The cv_bridge packages contains the CvBridge class which provides an interface between ROS and OpenCV. It helps in conversion from ROS Image messages to OpenCV images and vice-versa. You can learn more about it here(3)

Now let’s go over the instructions to install these tools one-by-one :

Installation of virtualenvwrapper

We will use Python’s Official package manager, pip for installing virtualenvwrapper. Open a terminal and enter the following command :

pip install virtualenvwrapper

Now that we have virtualenvwrapper downloaded and installed, we need to add it’s path in the .bashrc file so that your Linux installation will be able to locate it.

The .bashrc file contains terminal configurations, commands, environment variables etc. You can learn more about this here

Follow the steps given below for the same :

nano ~/.bashrc

Paste these lines in the .bashrc file which you opened above:

export WORKON_HOME=$HOME/.virtualenvs
export PROJECT_HOME=$HOME/Devel
source /usr/local/bin/virtualenvwrapper.sh

Now that we have included the paths in the .bashrc file, we need to execute the given command to load the new environment variables :

source ~/.bashrc

If you face an error here, then reopen the .bashrc file by :

nano ~/.bashrc

Now change the source :

source /usr/local/bin/virtualenvwrapper.sh

to :

source $HOME/.local/bin/virtualenvwrapper.sh

Again try loading the new environment variables by :

source ~/.bashrc

Creating a New Virtual Environment

Now that we have virtualenvwrapper locked and loaded, let’s create a new virtual environment.

The command given below will be used to create a new virtual environment :

mkvirtualenv [name of your virtual environment]

To use the new virtual environment

workon [name of your virtual environment]

To deactivate virtual environment

deactivate

Installation of OpenCV

Follow the steps given below to install OpenCV. We will again use pip for downloading and installing OpenCV.

pip install opencv-python

To test our OpenCV installation, Open Python interpreter in terminal by :

python

Let’s try importing OpenCV in the above opened Python interpreter

import cv2

If your import statement executes error-free, you are good to go !!

Installation of cv_bridge

Follow these steps to install cv_bridge :

Open a new terminal and execute the given command

sudo apt-get install ros-noetic-cv-bridge

Execute the command given below to confirm your cv_bridge installation :

rospack find cv_bridge

Note

cv_bridge is a ros package, so make sure you source your ROS installation before running the command below.