Updated December 21, 2023

Introduction to OpenCV on Ubuntu

Open-source computer vision (OpenCV) is a publicly accessible artificial intelligence and machine learning software library for computer vision applications. OpenCV adapts to multiple programming platforms and offers various tools and algorithms. It aligns best with segments like image processing, object detection, motion tracking, facial recognition, etc. Deploying OpenCV on Ubuntu allows smooth functioning and high performance.

Installation and Implementation are quite convenient using the Ubuntu terminal. Let us dive into its installation process.

Key takeaway

OpenCV integrates with various programming languages.
OpenCV is a versatile library comprising over 2500 algorithms and countless functionalities for computer vision tasks.
OpenCV facilitates parallel processing and computing, allowing for superior performance.
It is easily integrated with the scientific computational library for influential vision projects.
Easy installation and Implementation.

Prerequisites

Installed Ubuntu system with version 20.04 or 22.04 and a good internet connection
Essential command skills and terminal access
A user account with administrative privileges
Updated and upgraded system
Installed development tools and dependencies.

Note– All the above prerequisites are discussed below.

Update and upgrade system packages

Ensure the system is updated and upgraded with the latest packages before installing OpenCV to avoid conflict. To update and upgrade the system, use the following command line. Search for “terminal” to open a command prompt.

Update command

sudo apt update

It is good practice to use the update command before every software installation.

Upgrade

sudo apt upgrade

Install Dependencies

We must install the essential development tools, image and video processing libraries, Python packages, and other dependencies to set up OpenCV and ensure its smooth functioning. Below is the command line to install the essential dependencies

1) Basic Development Tools

sudo apt install build-essential cmake

2) Graphical user interface libraries

sudo apt install libgtk-3-dev

3) Audio and video processing and support

Library for Audio and Video Codes:

sudo apt install
libavcodec-dev
libavformat-dev
libswscale-dev

Library for Video4Linux and capture support

sudo apt install libv4l-dev

Library for Xvid and x264 video codecs

sudo apt install libxvidcore-dev libx264-dev

4) (HDR) Image Format:

sudo apt install openexr libopenexr-dev

5) Algebra and Mathematical Optimization:

sudo apt install libatlas-base-dev gfortran

6) Multimedia Libraries:

sudo apt install libgstreamer-plugins-base1.0-dev libgstreamer1.0-dev

7) Python libraries

For Integrating Python and OpenCV

sudo apt install python3-dev

Allow package management with pip and set virtual environment.

sudo apt install python3-pip
sudo apt install python3-venv

To perform Python numerical operations with OpenCV

pip install numpy

8) Threading Building Blocks library

sudo apt install libtbb
2 libtbb-dev

9) Image formatting Libraries:

sudo apt install libjpeg-dev libpng-dev libtiff-dev

10) FireWire Camera Support:

sudo apt install libdc1394-dev

How to Install OpenCV on Ubuntu? (Step by Step)

Ubuntu allows us to install OpenCV with different methods, some of which are listed below:

Method 1: Installing OpenCV from Ubuntu Repositories

The easiest and most efficient way to install OpenCV is from the standard Ubuntu repository. Run the below command for easy installation. But 1^st, update the Ubuntu packages.

Step 1: Update the Ubuntu packages list

sudo apt update

Step 2: Install OpenCV

sudo apt install libopencv-dev

Use the apt package manager to install OpenCV. The command above will install the latest development library in the Ubuntu repository.

Step 3: Check the version to confirm the OpenCV installation by using the command:

dpkg -l libopencv-dev

As we can see, 4.6.0 is installed.

Method 2: Installing OpenCV on Ubuntu from the source.

This means we will make a folder to copy the OpenCV repo using git. This copying of the files is known as cloning, and after the cloning is done, we will install OpenCV.

Step 1: Let us start with our step of updating the Ubuntu.

sudo apt update

Step 2: Install essential libraries and plugins to build OpenCV, as explained in the “install dependencies” section.

Step 3: Create a build directory as “opencv_build” or any other convenient name you choose, and access the directory.

mkdir ~/opencv_build && cd ~/opencv_build

Step 4: clone the official OpenCV repository with git

git clone https://github.com/opencv/opencv.git

Step 5: We also need to copy some additional remaining modules of OpenCV

git clone https://github.com/opencv/opencv_contrib.git

Step 6: We need to build an additional folder inside the opencv_build folder and go inside that folder

mkdir -p ~/opencv_build/opencv/build && cd ~/opencv_build/opencv/build

Step 7: Now, we need to set the configuration of the OpenCV before building and installing it. In the below command, we will specify the type of build, installation of Python and C, prefixes required for installation, etc.

cmake -D CMAKE_BUILD_TYPE=Release

-D CMAKE_INSTALL_PREFIX=/usr/local

-D INSTALL_C_EXAMPLES=ON

-D INSTALL_PYTHON_EXAMPLES=ON

-D OPENCV_GENERATE_PKGCONFIG=ON

-D BUILD_EXAMPLES=ON

-D OPENCV_EXTRA_MODULES_PATH=~/opencv_build/opencv_contrib/modules ..

Step 8: Once all the configurations are defined, start to build.

make -j3

Step 9: Installing is the last step.

sudo make install

Method 3: Installation using Python

In this method, it is recommended to create an environment where the code is clean, there is no conflict between projects and packages, and the storing of dependencies is done separately. Such an environment is technically termed a virtual environment. This virtual environment can be achieved using the Python module
named “venv.”
Step 1: “venv” doesn’t come by default with Python, so we need to install it.

sudo apt install python3.10-venv –y

Step 2: Create a suitable environment with a proper name after the installation. In our case, it is “myenv.”

sudo python3 -m venv myenv

Step 3: In order to work under the environment, we need to activate the environment.

Source myenv/bin/activate

Step 4: Once we have entered the virtual environment, install OpenCV

sudo pip3 install opencv-python

Method 4: Installation using Anaconda.

Anaconda is a Python distribution mainly known for installing and managing packages and creating a package management system.

The installation algorithm is the same as the above method; only the commands differ.

Step 1: Set up an environment to work in.

Conda create -n myenv python=3.10.6

Step 2: Once the environment is created, activate it using the below command

conda activate myenv

Step 3: After entering into the environment, installation of OpenCV can be executed using the command

conda install -c conda-forge opencv

Troubleshooting Common Issues

We might face some issues and errors while installing OpenCV

Here are some common problems and techniques to encounter them

1. Missing dependencies

Issue: Errors during installation indicating the missing packages

Solution: look for the lost package from the error window and install that package. Use the below command.

sudo apt install <missing_package_name>

2. Dependencies package

Issue: For different projects, we require different versions of packages and these cause dependencies conflict

Solution: set up the virtual environment using the “venv” module of Python. This creates a package management system to avoid conflicts.

3. Permission issue

Issue: Permission is denied while installing a particular package.

Solution: Make sure to add “sudo” at the start of the command.

Example of Using OpenCV in Python

Explanation of crucial steps in the below program

Step 1: Import the OpenCV library and OS module to define the image path.

Step 2: Define a function to display and save the image: display_Python_image and save_Python_image

Step 3: Read the image with the cv2.imread function

Step 4: Check for the image; display an error message if it is not found.

Step 5: Display the original image with the display_Python_image function

Step6: perform different OpenCV functions on the image to alter the image and display them

Manipulate the image to black and white format with the cv2.cvtColor function
Change the dimensions of the image with the cv2.resize function
Detect image edges with cv2.Canny function
Rotate the image to a certain angle with the cv2.getRotationMatrix2D function.

import cv2
import os

def display_Python_image(window_name, image):
    cv2.imshow(window_name, image)
    cv2.waitKey(0)
    cv2.destroyAllWindows()

def save_Python_image(image, output_path):
    cv2.imwrite(output_path, image)
    print(f"Image is saved at {output_path}")

# Read Python image from the file
script_directory = os.path.dirname(os.path.realpath(__file__))
Python_image_path = os.path.join(script_directory, "Python_image.jpg")
original_Python_image = cv2.imread(Python_image_path)

# load the image
if original_Python_image is None:
    print(f"Image_Error: Unable to load the original image from {Python_image_path}")
else:
    # Displaying the original image
    display_Python_image("Displaying Original Image", original_Python_image)

    # Converting image to Black and White color
    BlackWhite_image = cv2.cvtColor(original_Python_image, cv2.COLOR_BGR2GRAY)

    # Displaying the Black and White image
    display_Python_image("Displaying image as balck and white", BlackWhite_image)

    # Save Black and White image
    BlackWhite_output_path = "path/to/your/BlackWhite_image.jpg"
    save_Python_image(BlackWhite_image, BlackWhite_output_path)

    # Resizing Image
    resized_pythonimage = cv2.resize(original_Python_image, (450, 450))
    display_Python_image("image with new dimension", resized_pythonimage)

    # Edge Detection
    image_edges = cv2.Canny(original_Python_image, 100, 200)
    display_Python_image("Detection of picture edges", image_edges)

    # Rotating Image
    rotation_matrix = cv2.getRotationMatrix2D((original_Python_image.shape[1] // 2, original_Python_image.shape[0] // 2), 45, 1)
    rotated_pythonimage = cv2.warpAffine(original_Python_image, rotation_matrix, (original_Python_image.shape[1], original_Python_image.shape[0]))
    display_Python_image("Rotated python Image", rotated_pythonimage)

    # Save Rotate Image
    rotated_image_route = "path/to/your/rotated_pythonimage.jpg"
    save_Python_image(rotated_pythonimage, rotated_image_route)

Output:

Displaying original image:

Displaying Black and White image:

Displaying image with new dimensions:

Detecting image Edges:

Rotated Image:

Conclusion

OpenCV allows developers to unlock robust computer vision features and create superior applications. It offers flexible methods to install virtually according to the developer’s requirements. With the help of an Ubuntu terminal and simple commands, OpenCV can be installed using four methods (repository, git, Anaconda, pip ), as discussed above.

FAQs

Q 1. How to integrate OpenCV with NumPy, TensorFlow, and PyTorch library

Answer: Integrating OpenCV with these libraries allows a smooth process for performing computing tasks. Below is a step-by-step guideline for library integration
Step 1: Import essential libraries for a specific task to be performed
Step2: Pre-process image with OpenCV functions like read, resize, and format image
Step3: Choose the appropriate model architecture from these computational libraries to define a mode
Step4: Train model with labeled datasets of images and perform data manipulation with OpenCV functions
Step 5: Make predictions and perform image analysis

Q2. How to manage multiple OpenCV versions on my OS

Answer: Some ways to manage various versions are:

Isolated environment: Make a virtual environment separately for each specific OpenCV project
Package manager: To manage different versions, use conda, pipenv, venv, pyenv, and virtualenvwrapper tools
Installing manually: Make installation in separate directories and paths.

Q 3. How do you encounter dependency issues on Ubuntu?

Answer: To resolve the dependency issue, follow some steps:
1) Recognize missing package: In the error message, check the required dependency. To check package dependency, run the below command:
apt show [package name] or dpkg -I [package file] 2) Keep the package list up to date by below command
sudo apt update
3) Install missing dependencies. There are two ways to install dependencies:

Automatic installation:
sudo apt install [package name] -f

Manual installation
sudo apt install [dependency name]

Quiz Result
Total Questions	Correct Answers	Wrong Answers	Percentage

Introduction to OpenCV on Ubuntu

Table of contents:

Key takeaway

Prerequisites

Update and upgrade system packages

Install Dependencies

How to Install OpenCV on Ubuntu? (Step by Step)

Method 1: Installing OpenCV from Ubuntu Repositories

Method 2: Installing OpenCV on Ubuntu from the source.

Method 3: Installation using Python

Method 4: Installation using Anaconda.

Troubleshooting Common Issues

1. Missing dependencies

2. Dependencies package

3. Permission issue

Example of Using OpenCV in Python

Conclusion

FAQs

Recommended Articles

Follow us!

APPS

Blog

Courses

Email

OpenCV Ubuntu

Introduction to OpenCV on Ubuntu

Table of contents:

Key takeaway

Prerequisites

Update and upgrade system packages

Install Dependencies

How to Install OpenCV on Ubuntu? (Step by Step)

Method 1: Installing OpenCV from Ubuntu Repositories

Method 2: Installing OpenCV on Ubuntu from the source.

Method 3: Installation using Python

Method 4: Installation using Anaconda.

Troubleshooting Common Issues

1. Missing dependencies

2. Dependencies package

3. Permission issue

Example of Using OpenCV in Python

Conclusion

FAQs

Recommended Articles