PyImageSearch

You can master Computer Vision, Deep Learning, and OpenCV - PyImageSearch

OpenCV Tutorials Tutorials

Ubuntu 18.04: How to install OpenCV

by on

In this blog post you will learn how to install OpenCV on Ubuntu 18.04.

In the past, I’ve authored a handful of installation guides for Ubuntu:

The folks at Canonical have been working hard.

On April 26, 2018, they’ve released a new Long Term Support (LTS) version of Ubuntu for the community: Ubuntu 18.04 LTS (Bionic Beaver).

Support for Ubuntu 16.04 LTS continues until April 2021 so rest-assured — you don’t have to upgrade your 16.04 OS to continue working on your image processing projects.

That said, if you want to upgrade to Ubuntu 18.04 and use the latest-and-greatest, I think you’ll be quite pleased with the new changes in Ubuntu 18.04.

Let’s get down to business and install OpenCV with Python 3 bindings.

To learn how to stand up your Ubuntu 18.04 system with OpenCV, just keep reading.

Note: While you won’t see an Ubuntu 17.10 specific guide here on my blog (non-LTS), these instructions may work with 17.10 (you’ll just have to proceed at your own risk).

Ubuntu 18.04: How to install OpenCV

One major change in Ubuntu 18.04 is that they’ve dropped Python 2.7 completely.

You can still install Python 2.7 if-needed, but now Python 3 is the default on the OS.

Given that, this guide supports Python 3. If you need Python 2.7 support, read this entire guide first and then check the first question of the Troubleshooting your install (FAQ) section near the bottom of this blog post for some Python 2.7 pointers.

Step #0: Get comfortable — you’ll be using Python 3.6

Let’s familiarize ourselves with Python 3 on Ubuntu 18.04.

To run Python 3 on Ubuntu 18.04, you must call python3 explicitly.

Let’s see which version is installed on our system:

$ python3 --version
Python 3.6.5

And now, let’s launch a Python 3 shell just to test the waters:

$ python3
>> print("OpenCV + Ubuntu 18.04!")
OpenCV + Ubuntu 18.04!
>> quit()

That’s easy enough, so let’s get on with installing OpenCV on Ubuntu 18.04.

Step #1: Install OpenCV dependencies on Ubuntu 18.04

All steps today will be accomplished in the terminal/command line. Before we begin, open a terminal or connect via SSH.

From there, we need to refresh/upgrade the pre-installed packages/libraries with the apt-get package manager:

$ sudo apt-get update
$ sudo apt-get upgrade

Followed by installing developer tools:

$ sudo apt-get install build-essential cmake unzip pkg-config

You most likely already have pkg-config installed on Ubuntu 18.04, but be sure to include it in the install command for sanity.

Next, we need to install some OpenCV-specific prerequisites. OpenCV is an image processing/computer vision library and therefore it needs to be able to load standard image file formats such as JPEG, PNG, TIFF, etc. The following image I/O packages will allow OpenCV to work with image files:

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

Now let’s try to install libjasper-dev :

$ sudo apt-get install libjasper-dev

If you receive an error about libjasper-dev being missing then follow the following instructions:

sudo add-apt-repository "deb http://security.ubuntu.com/ubuntu xenial-security main"
sudo apt update
sudo apt install libjasper1 libjasper-dev

Otherwise (or once libjasper-dev is installed), keep going.

Next, let’s include video I/O packages as we often work with video on the PyImageSearch blog. You’ll need the following packages so you can work with your camera stream and process video files:

$ sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libv4l-dev
$ sudo apt-get install libxvidcore-dev libx264-dev

OpenCV’s highgui module relies on the GTK library for GUI operations. The highgui module will allow you to create elementary GUIs which display images, handle kepresses/mouse clicks, and create sliders and trackbars. Advanced GUIs should be built with TK, Wx, or QT. See this blog post to learn how to make an OpenCV GUI with TK.

Let’s install GTK:

$ sudo apt-get install libgtk-3-dev

I always recommend the following two libraries which will optimize various OpenCV functions:

$ sudo apt-get install libatlas-base-dev gfortran

And finally, our last requirement is to install Python 3 headers and libraries:

$ sudo apt-get install python3.6-dev

Step #2: Download the official OpenCV source

Update 2018-12-20: These instructions have been updated to work with OpenCV 3.4.4. These instructions should continue to work with future OpenCV 3.x versions as well.

Since we’re continuing to work in the terminal, let’s download the official OpenCV release using wget :

$ cd ~
$ wget -O opencv.zip https://github.com/opencv/opencv/archive/3.4.4.zip

Followed by the opencv_contrib module:

$ wget -O opencv_contrib.zip https://github.com/opencv/opencv_contrib/archive/3.4.4.zip

Note: If your browser is cutting off the full command either use the “<>” button in the toolbar above to expand the code block or copy and paste the following URL: https://github.com/opencv/opencv_contrib/archive/3.4.4.zip

So, what’s the contrib repo?

The contrib repository contains algorithms such as SIFT, SURF, and others. In the past, these implementations were included in the default installation of OpenCV 2.4; however, they were moved beginning with OpenCV 3+.

Modules that are actively being developed and/or modules that are patented (not free for commercial/industry use) are included in the contrib module. SIFT and SURF fall into this category. You can learn more about the thought process behind this move in the following blog post: Where did SIFT and SURF go in OpenCV 3?

Important: Both opencv and opencv_contrib versions must be identical. Notice that both URLs point to 3.4.4. Feel free to install a different version while still using this guide — just be sure to update both URLs.

Now, let’s unzip the archives:

$ unzip opencv.zip
$ unzip opencv_contrib.zip
Figure 1: After downloading and unzipping opencv and opencv_contrib, our home directory listing should look similar to what is displayed in the terminal.

Now let’s go ahead and rename the directories:

$ mv opencv-3.4.4 opencv
$ mv opencv_contrib-3.4.4 opencv_contrib

Step #3: Configure your Python 3 environment

The first step we’re taking to configure our Python 3 development environment is to install pip, a Python Package Manager.

To install pip, simply enter the following in your terminal:

$ wget https://bootstrap.pypa.io/get-pip.py
$ sudo python3 get-pip.py

Making use of virtual environments for Python development

If you are familiar with my blog and install guides therein, the following statement might make me sound like a broken record but I’ll repeat it anyway:

I use both virtualenv and virtualenvwrapper daily and you should too unless you have a very specific reason not to. These two Python packages facilitate creating independent Python environments for your projects.

It is a best practice to use virtual environments.

Why?

Virtual environments allow you to work on your projects in isolation without spinning up resource hogs such as VMs and Docker images (I definitely do use both VirtualBox and Docker — they have their place).

For example, maybe you have a Python + OpenCV project that requires an older version of scikit-learn (v0.14) but you want to keep using the latest version of scikit-learn (0.19) for all of your newer projects.

Using virtual environments, you could handle these two software version dependencies separately, something that is not possible using just the system install of Python.

If you would like more information about Python virtual environments take a look at this article on RealPython or read the first half of the this blog post on PyImageSearch.

Let’s go ahead and install virtualenv and virtualenvwrapper now:

$ sudo pip install virtualenv virtualenvwrapper
$ sudo rm -rf ~/get-pip.py ~/.cache/pip

To finish the install we need to update our ~/.bashrc file.

Using a terminal text editor such as vi /vim or nano , add the following lines to your ~/.bashrc :

# virtualenv and virtualenvwrapper
export WORKON_HOME=$HOME/.virtualenvs
export VIRTUALENVWRAPPER_PYTHON=/usr/bin/python3
source /usr/local/bin/virtualenvwrapper.sh

Alternatively, you can append the lines directly via bash commands:

$ echo -e "\n# virtualenv and virtualenvwrapper" >> ~/.bashrc
$ echo "export WORKON_HOME=$HOME/.virtualenvs" >> ~/.bashrc
$ echo "export VIRTUALENVWRAPPER_PYTHON=/usr/bin/python3" >> ~/.bashrc
$ echo "source /usr/local/bin/virtualenvwrapper.sh" >> ~/.bashrc

Next, source the ~/.bashrc file:

$ source ~/.bashrc

Creating a virtual environment to hold OpenCV and additional packages

Ok, while that may have seemed like a lot of work, we’re at the point where we can create your Python 3 virtual environment for OpenCV:

$ mkvirtualenv cv -p python3

This line simply creates a Python 3 virtual environment named cv . You can name your environment(s) whatever you’d like — I like to keep them short and sweet while also providing enough information so I’ll remember what they are for. You can have as many virtual environments on your system as you’d like!

Let’s verify that we’re in the cv environment by using the workon command:

$ workon cv

Figure 2 shows what your terminal will look like (assuming you haven’t changed any bash prompt settings):

Figure 2: If you see the (cv) at the beginning of the bash prompt, then your virtual environment is active and you’re working “inside” the environment. You can now safely install OpenCV with correct Python bindings.

Install NumPy in your environment

Let’s install our first package into the environment: NumPy. NumPy is a requirement for working with Python and OpenCV. We simply use pip (while the cv Python virtual environment is active):

$ pip install numpy

Step #4: Configure and compile OpenCV for Ubuntu 18.04

Now we’re moving. We’re ready to compile and install OpenCV.

Before we begin though, let’s ensure that we’re in the cv virtual environment:

$ workon cv

It is very important that the virtual environment is active (you are “inside” the virtual environment) which is why I keep reiterating it. If you are not in the cv Python virtual environment before moving on to the next step your build files will not be generated properly.

Configure OpenCV with CMake

Let’s set up our OpenCV build using cmake :

$ cd ~/opencv
$ mkdir build
$ cd build
$ cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D WITH_CUDA=OFF \
    -D INSTALL_PYTHON_EXAMPLES=ON \
    -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib/modules \
    -D OPENCV_ENABLE_NONFREE=ON \
    -D BUILD_EXAMPLES=ON ..

Update 2018-12-20: Be sure to set the OPENCV_ENABLE_NONFREE=ON flag so you’ll have access to patented algorithms that we sometimes review on this blog (and cover in depth in PyImageSearch Gurus). Likewise, be sure to update the path to your virtual environment so that the correct PYTHON_EXECUTABLE will be used (notice cv in the path since our virtual environment name is cv ).

I always recommend that you scroll through the CMake output and check to see if anything looks out of the ordinary. You won’t see a “YES” marked next to every setting — that is normal. Be sure you don’t see any errors or your compile may fail (warnings are okay).

Figure 3: To compile OpenCV for Ubuntu 18.04, we make use of CMake. The CMake tool will configure settings prior to compilation.

Take a moment to notice that only “Python 3” section is shown in the CMake output on Ubuntu 18.04 in Figure 3. This is by design as we are only compiling OpenCV with Python 3 support.

Update 2018-12-20: Additionally, scroll up about 20 lines and make sure “Non-free algorithms” is set to “YES”.

Note: If you are encountering problems related to stdlib.h: No such file or directory during either the cmake or make phase of this tutorial you’ll also need to include the following option to CMake: -D ENABLE_PRECOMPILED_HEADERS=OFF . In this case I would suggest deleting your build directory, re-creating it, and then re-running cmake with the above option included. This will resolve the stdlib.h error.

Compiling OpenCV on Ubuntu 18.04

Let’s compile OpenCV using make .

Depending on the number of processors/cores, you may be able to reduce compile time by altering the flag in the command. My computer has 4 cores, so I am using the -j4 flag. You can update the numeral or leave the flag off altogether:

$ make -j4
Figure 4: To compile OpenCV with Python 3 on Ubuntu 18.04, we use make. Using make compiles OpenCV from source and is preferred over using package managers for installing OpenCV.

This process may take 30 minutes or longer, so go for a nice walk if you are able.

If your compile chokes and hangs, it may be due to a threading race condition. In the event you run into this problem, simply delete your build directory, recreate it, and re-run cmake and make . This time do not include the flag next to make .

Installing and verifying OpenCV

Upon a successful, 100% complete compile you can now install OpenCV:

$ sudo make install
$ sudo ldconfig

To verify the install, sometimes I like to enter the following command in the terminal:

$ pkg-config --modversion opencv
3.4.4

Step #5: Finish your Python+ OpenCV + Ubuntu 18.04 install

We’ve reached the last lap of the race so stick with it.

Update 2018-12-20: The following paths have been updated. Previous versions of OpenCV installed the bindings in a different location (/usr/local/lib/python3.6/site-packages ), so be sure to take a look at the paths below carefully.

At this point, your Python 3 bindings for OpenCV should reside in the following folder:

$ ls /usr/local/python/cv2/python-3.6
cv2.cpython-36m-x86_64-linux-gnu.so

Let’s rename them to simply cv2.so :

$ cd /usr/local/python/cv2/python-3.6
$ sudo mv cv2.cpython-36m-x86_64-linux-gnu.so cv2.so

Pro-tip: If you are installing OpenCV 3 and OpenCV 4 alongside each other, instead of renaming the file to cv2.so, you might consider naming it cv2.opencv3.4.4.so and then in the next sub-step sym-link appropriately from that file to cv2.so as well.

Our last sub-step is to sym-link our OpenCV cv2.so bindings into our cv virtual environment:

$ cd ~/.virtualenvs/cv/lib/python3.6/site-packages/
$ ln -s /usr/local/python/cv2/python-3.6/cv2.so cv2.so

Step #6: Testing your OpenCV 3 install on Ubuntu 18.04

The race is done, but let’s verify that we’re firing on all cylinders.

To verify that our OpenCV + Ubuntu install is complete, I like to launch Python, import OpenCV, and query for the version (this is useful for sanity if you have multiple versions of OpenCV installed as well):

$ cd ~
$ workon cv
$ python
Python 3.6.5 (default, Apr 1 2018, 05:46:30)
[GCC 7.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> cv2.__version__
'3.4.4'
>>> quit()

Here’s what it looks like on my system:

Figure 5: To verify that OpenCV is correctly installed and configured in our Python 3 virtual environment, I like to run the Python interpreter in the terminal. From there you can import OpenCV (cv2) and verify the version number matches what you intended to install.

Update 2018-12-20: I did not update the screenshot today during my update. The version should reflect the version of OpenCV you installed.

Optionally, at this point, you can safely delete the zips and directories in your home folder:

$ cd ~
$ rm opencv.zip opencv_contrib.zip
$ rm -rf opencv opencv_contrib

Troubleshooting your install (FAQ)

In this section, I address some of the common questions, problems, and issues that arise when installing OpenCV 3 with Python 3 on Ubuntu 18.04 LTS.

Q. Where is Python 2.7 on Ubuntu 18.04?

A. Python 3 is the default and what comes with Ubuntu 18.04. Python 2.7 users can manually install Python 2.7 at the end of Step #1:

$ sudo apt-get install python2.7 python2.7-dev

From there, when you create your virtual environment in Step #3, first install pip for Python 2.7:

$ sudo python2.7 get-pip.py

And then (also in Step #3) when you make your virtual environment, simply use the relevant Python version flag:

$ mkvirtualenv cv -p python2.7

From there everything should be the same.

Q. Why can’t I just pip to install OpenCV?

A. There are a number of pip-installable versions of OpenCV available depending on your operating system and architecture. The problem you may run into is that they may be compiled without various optimizations, image I/O support, video I/O support, and opencv_contrib  support. Use them — but use them at your own risk. This tutorial is meant to give you the full install of OpenCV on Ubuntu 18.04 while giving you complete control over the compile.

Q. When I execute mkvirtualenv  or workon , I encounter a “command not found error”.

A. There a number of reasons why you would be seeing this error message, all of come from to Step #3:

  1. First, make sure you have installed virtualenv  and virtualenvwrapper  properly using the pip  package manager. Verify by running pip freeze , and ensure that you see both virtualenv  and virtualenvwrapper  in the list of installed packages.
  2. Your ~/.bashrc  file may have mistakes. View the contents of your ~/.bashrc  file to see the proper export  and source  commands are present (check Step #3 for the commands that should be appended to ~/.bashrc ).
  3. You may have forgotten to source  your ~/.bashrc . Make sure you run source ~/.bashrc  after editing it to ensure you have access to the mkvirtualenv  and workon  commands.

Q. When I open a new terminal, logout, or reboot my Ubuntu system, I cannot execute the mkvirtualenv  or workon  commands.

A. See #2 from the previous question.

Q. When I try to import OpenCV, I encounter this message: Import Error: No module named cv2 .

A. There are multiple reasons this could be happening and unfortunately, it is hard to diagnose. I recommend the following suggestions to help diagnose and resolve the error:

  1. Make sure your cv  virtual environment is active by using the workon cv  command. If this command gives you an error, then see the first question in this FAQ.
  2. Try investigating the contents of the site-packages  directory in your cv  virtual environment. You can find the site-packages  directory in ~/.virtualenvs/cv/lib/python3.6/site-packages/  depending on your Python version. Make sure (1) there is a cv2.so  file in the site-packages  directory and (2) it’s properly sym-linked to a valid file.
  3. Be sure to check the site-packages  (and even dist-packages ) directory for the system install of Python located in /usr/local/lib/python3.6/site-packages/ . Additionally please check /usr/local/python/cv2 . Ideally, you should have a cv2.so  file in one of those locations.
  4. As a last resort, check in your build/lib  directory of your OpenCV build. There should be a cv2.so  file there (if both cmake  and make  executed without error). If the cv2.so  file is present, manually copy it into both the system site-packages  directory as well as the site-packages  directory for the cv  virtual environment.

What's next? We recommend PyImageSearch University.

Course information:
86 total classes • 115+ hours of on-demand code walkthrough videos • Last updated: October 2024
★★★★★ 4.84 (128 Ratings) • 16,000+ Students Enrolled

I strongly believe that if you had the right teacher you could master computer vision and deep learning.

Do you think learning computer vision and deep learning has to be time-consuming, overwhelming, and complicated? Or has to involve complex mathematics and equations? Or requires a degree in computer science?

That’s not the case.

All you need to master computer vision and deep learning is for someone to explain things to you in simple, intuitive terms. And that’s exactly what I do. My mission is to change education and how complex Artificial Intelligence topics are taught.

If you're serious about learning computer vision, your next stop should be PyImageSearch University, the most comprehensive computer vision, deep learning, and OpenCV course online today. Here you’ll learn how to successfully and confidently apply computer vision to your work, research, and projects. Join me in computer vision mastery.

Inside PyImageSearch University you'll find:

  • ✓ 86 courses on essential computer vision, deep learning, and OpenCV topics
  • ✓ 86 Certificates of Completion
  • ✓ 115+ hours of on-demand video
  • ✓ Brand new courses released regularly, ensuring you can keep up with state-of-the-art techniques
  • ✓ Pre-configured Jupyter Notebooks in Google Colab
  • ✓ Run all code examples in your web browser — works on Windows, macOS, and Linux (no dev environment configuration required!)
  • ✓ Access to centralized code repos for all 540+ tutorials on PyImageSearch
  • ✓ Easy one-click downloads for code, datasets, pre-trained models, etc.
  • ✓ Access on mobile, laptop, desktop, etc.

Click here to join PyImageSearch University

Summary

Today we installed OpenCV 3 with Python 3 bindings on Ubuntu 18.04 LTS.

I hope you found these instructions helpful on getting your own Ubuntu 18.04 machine configured with OpenCV 3.

If you’re interested in learning more about OpenCV, computer vision, and image processing, be sure to enter your email address in the form below to be notified when new blog posts + tutorials are published!

Join the PyImageSearch Newsletter and Grab My FREE 17-page Resource Guide PDF

Enter your email address below to join the PyImageSearch Newsletter and download my FREE 17-page Resource Guide PDF on Computer Vision, OpenCV, and Deep Learning.

About the Author

Hi there, I’m Adrian Rosebrock, PhD. All too often I see developers, students, and researchers wasting their time, studying the wrong things, and generally struggling to get started with Computer Vision, Deep Learning, and OpenCV. I created this website to show you what I believe is the best possible way to get your start.

Reader Interactions

Previous Article:

An OpenCV barcode and QR code scanner with ZBar

Next Article:

Keras: Multiple outputs and multiple losses

194 responses to: Ubuntu 18.04: How to install OpenCV

  1. El Barto

    Hi Adrian! Is there a certain reason why you don’t clone the OpenCV repo instead of fetching as an archive?

    • Adrian Rosebrock

      You could do either, but downloading the official release archive tends to be easier for those who are not used to installing and configuring OpenCV on their systems (less commands to execute hear typically means less chance of causing a problem). If you cloned down the original repo the download would take longer and you would have the added step of checking out the specific release.

  2. Sarkar

    Hey Adrian, thanks for another awesome guide. Is it possible to use conda (instead of pip) in the above guide?

    thanks

    • Adrian Rosebrock

      You should be able to but I haven’t tested with Anaconda (I’m not an Anaconda user). The most important part will be during the “cmake” step where you’ll want to make sure your “Python 3” libraries, interpreter, etc. point to your Anaconda version, not your system version — this is a mistake I’ve seen others make.

      • Anthony Park

        Hi, Adrian.
        Could you be more specific about the cmake step for Anaconda users? I followed your kind instruction and compiled successfully, but I don’t see site-packages created and assuming this is because I have Anaconda version of python. How can I appropriately make sure python 3 libraries, etc point to my Anaconda version?
        Thank you 🙂

        • Adrian Rosebrock

          Hey Anthony — I don’t have any systems with Anaconda installed so I unfortunately cannot provide you with an example cmake command as I do not have the directory structure you would. The closest example I would have is this somewhat related tutorial on installing OpenCV on macOS. Inside the tutorial you’ll see that I set the PYTHON3_LIBRARY and PYTHON3_INCLUDE_DIR values. You will need to do the same for Anaconda.

  3. Adam

    Hey Adrian, thanks a lot for another awesome tutorial!

    1) Are you planning to release another one for the lucky GPU owners? 😉
    2) Are you aware of any compatibility issues for those who are into your Deep Leaning book? 😉

    Cheers,
    Adam

    • Adrian Rosebrock

      1. I’m not sure what you mean. Are you asking for specific deep learning library install instructions?

      2. I am not aware of any compatibility issues but I also haven’t performed exhaustive experiments on Ubuntu 18.04 either.

      • Adam

        1. Yes. I was wondering if the instructions for Ubuntu 18.04 deep learning installation will be similar to 16.04. It took me literally weeks to get all lib versions (TF, cuDNN, Cuda) to work together on 16.04. It was a pain and I failed miserably when tried to wrap it all inside a clean Docker image ;/

        2. That’s cool Adrian

        • Adrian Rosebrock

          When it comes to deep learning environments, if it’s not broken, don’t fix it. Ubuntu 16.04 is perfectly legitimate for deep learning work. I would not be chomping at the bit to immediately switch to Ubuntu 18.04 for deep learning unless there was a very specific reason for doing so.

  4. Thúlio

    Hello Adrian! I have a question about the new Ubuntu release. Have you been using Ubuntu 18.04 for deep learning? I’ve installed it and I’m planning on using TensorFlow and MXNet with GPU support enabled, but I’m afraid I was too eager to try the new release and didn’t think about the lack of support for those packages since 18.04 is too new. Should I have sticked with 16.04? Or are TF and MXNet already good to go on 18.04?

    • Adrian Rosebrock

      As far as deep learning goes I would recommend sticking with Ubuntu 16.04 for the time being. Ubuntu 16.04 is more “mature” in terms of support of deep learning libraries. Keep in mind that Ubuntu 18.04 is essentially brand new and would be more likely of encountering issues. Deep learning libraries change rapidly though so I wouldn’t get too concerned about it if you’re already on 18.04.

  6. Doug

    This article is excellent as always and timely for me, as I spent the last few days making opencv3.4.1 with dnn_modern and python 2.7 and 3.6 libs in a ubuntu16.04 docker. The long build time you noted above – combined with my copious mistakes insured many joyous repetitions of the cmake/make bits, until I got the opencv cmake build options just right. (And I needed? to build python cv libs and load them before I could see if they could find all their shared libs, in turn.) I’ve tried virtualenv but it was just too easy for me to break my ubuntu desktop with it. The docker approach seems to make up for my sloppiness with virtualenv. I often need a bib when I eat, though, so that might be a pattern..

    • Adrian Rosebrock

      Congrats on getting OpenCV installed on your Ubuntu system, Doug! I agree that Docker is a good use case here. Your rational for finding it too easy to break your main system is part of the reason why I really like cloud-based systems. Configure it once, snapshot it, and then delete + spin up a new instance if you ever break it!

    • Adrian Rosebrock

      Those optimizations are for ARM processors. Most Ubuntu users will likely be on non-ARM chipsets. If you are installing Ubuntu + OpenCV on an ARM processor you can of course use them.

  8. Ben

    Hi Adrian,

    I am new to your blog but I have been working with success so far on the recent steps 1 & 2 of the Pokemon project (microsoft bing search & keras deep learning)…

    Everything seems to be working for me with Anaconda 3.6 and Linux Mint 18.3 Cinnamon 64 bit with installing/using the deep learning libraries..

    Im also new to Linux OS and an IT person recommended Mint to me for a place to start…

    Would you recommend at all changing to Ubuntu at some point in time? Is there anything special about Linux Mint and using Anaconda? At some point in time I may actually have enough experience to require virtual environments with different versions of software…

    Thanks much… Cheers.
    Ben

    • Adrian Rosebrock

      I’ll be honest, I don’t know much about Linux Mint so I don’t think I’m qualified to say whether you should be using Ubuntu over Linux Mint. That said, I Ubuntu is often used for many deep learning projects and it’s very user friendly. I would recommend it if you are using Linux systems.

  9. Chris

    Thanks, Adrian. This is a very good tutorial for me to implement opencv on my new computer. I have a question now, how can I work in jupyter notebook and use this cv virtual environment. Thanks very much.

    • Adrian Rosebrock

      You would need to install Jupyter into the “cv” Python virtual environment:

      $ workon cv
$ pip install jupyter

  10. Antony Smith

    Hi Adrian,
    I bapassed the virtual env installation, but upon testing the installation (after all else installing correctly), I’m getting the Error:

    ImportError: /usr/local/lib/python3.6/site-packages/cv2.so: undefined symbol: _ZTIN2cv3dnn19experimental_dnn_v35LayerE

    • Antony Smith

      Okay nevermind, my bad.
      Just opened a new terminal and retried and all is right with the world…
      Thanks man!
      You’re the best!

  11. Andres

    hi Adrian

    Iḿ having trouble sinvce I can find my site-packages folder in the path you say, there is only the dist-packages directory in that path. I don’t know where to look for it or if it is suposed to be only there.

    thank you in advance

    Andres

    • Adrian Rosebrock

      I’m not sure why you would not have a site-packages directory and only a dist-packages directory. They should both be in the same subdirectory. Which step are you on?

      • Andres

        I’m on step 5 making the link, all the make an make install has finished without error

        • Adrian Rosebrock

          If the “make install” worked you should check your “dist-packages” directory for the “cv2.so” bindings instead.

          • Chemmy

            I had the same problem on a clean VM.

            Try:

            sudo apt-get install python3-dev libpython3.6-dev python3-numpy

            Delete you build directory and make it again from step 4. You will have the site-packages folder after this and the cv2.cpython-36m-x86_64-linux-gnu.so file.

            F.Y.I. I have no idea what I am doing. I just googled the answer and this worked.

          • AlexSeo

            I solved by:
            – creating manually an empty directory called “site-packages” into “/usr/local/lib/python3.6/”
            – deleting “build” directory and starting again from step 4.

    • Jarek

      I have the same with OpenCv 3.4.4. It appears that the cv2.cpython-36m-x86_64-linux-gnu.so file was created in /usr/local/python/cv2/python-3.6

  12. Andres

    Hi Adrian

    I have changed to Ubuntu 16.04 since im going tu use DNN and it seems is more reliable in that version, I have a doubt about CUDA integration with my project because I’m having trouble integrating the library in opencv, since im going to use tensorflow which one is better to compile CUDA with? or i will need it with booth of them?

    • Adrian Rosebrock

      You can use OpenCV with CUDA but it really depends on what you are doing. If your goal is to be developing DNNs you should install TensorFlow/Keras with CUDA support. OpenCV can be used to make predictions with some DNNs but for training you should be using TensorFlow/Keras.

  13. Anand Setlur

    Hi Adrian, I had to modify your cmake script to get it to work with CUDA9.0 in Ubuntu18.04 to use gcc-6 and fix some other build errors by specifying the CUDA_NVCC_FLAGS and PYTHON2_EXECUTABLE options

    export CC=/usr/bin/gcc-6
    export CXX=/usr/bin/g++-6
    mkdir -p build
    cd build
    cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CUDA_NVCC_FLAGS=–expt-relaxed-constexpr \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D INSTALL_PYTHON_EXAMPLES=ON \
    -D INSTALL_C_EXAMPLES=OFF \
    -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.4.1/modules \
    -D PYTHON_EXECUTABLE=~/.virtualenvs/cv3/bin/python \
    -D PYTHON2_EXECUTABLE=/usr/bin/python2 \
    -D BUILD_EXAMPLES=ON ..
    make -j4

    • Robert

      Thanks Anand! I was still getting errors with your example above until I changed the CUDA_NVCC_FLAGS to:
      -D CUDA_NVCC_FLAGS=”-D_FORCE_INLINES –expt-relaxed-constexpr” \

      • Robert

        This is the complete command I used:

        cmake -D CMAKE_BUILD_TYPE=RELEASE \
        -D CUDA_NVCC_FLAGS=”-D_FORCE_INLINES –expt-relaxed-constexpr” \
        -D CMAKE_INSTALL_PREFIX=/usr/local \
        -D INSTALL_PYTHON_EXAMPLES=ON \
        -D INSTALL_C_EXAMPLES=OFF \
        -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.4.1/modules \
        -D PYTHON_EXECUTABLE=~/.virtualenvs/cv/bin/python \
        -D BUILD_EXAMPLES=ON ..

  14. artnect

    Hi Adrian

    I have gotten several failed process on building open cv using cmake

    can you help me on that ?

    the log file is too long i can’t post it here.

  15. Skanda Kumar

    Hi Adrian, I am a novice user of OpenCV. My supervisor at my internship was very specific about installing OpenCV version 2.4. I installed the latest version of Ubuntu(18.04) in my system and I am trying to install the the required version of OpenCV but I am unable to. This was the most detailed post I could find on the internet and I would be grateful if you can help with getting OpenCV 2.4 running in my Ubuntu 18.04. Thanks in advance 🙂

    • Adrian Rosebrock

      Oh man, I imagine trying to install OpenCV 2.4 on Ubuntu 18.04 would be quite a pain. It’s a hack but you may want to consider installing a VM with a much older version of Ubuntu 10.04/12.04 and then following the instructions from a previous previous Raspbian tutorial (Ubuntu and Raspbian are both Debian based). I do not know offhand what changes would need to be made for this tutorial and OpenCV 2.4 but I imagine there would be a few, mostly related to the specific apt-get packages installed. Best of luck with it!

    • clamytoe

      Might be too late now, but if you use Anaconda, you can install OpenCV 2.4.2. It probably won’t have all of the bells and whistles like this full installation, but it should get you going.

  16. Max Lee

    After following instructions, to specify here, I found that contrib modules, while being built, weren’t actually getting installed… after digging around, found that you need to modify the cmake command where the extra modules are specified, to add a target for them.

    Updated cmake command:
    cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D INSTALL_PYTHON_EXAMPLES=ON \
    -D INSTALL_C_EXAMPLES=OFF \
    -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.4.1/modules ~/opencv-3.4.1\
    -D PYTHON_EXECUTABLE=~/.virtualenvs/cv/bin/python \
    -D BUILD_EXAMPLES=ON ..

    This then allowed me to use the contrib modules for a few things that I was also compiling in C++.

  17. Josef Matondang

    Hello Adrian, do you know how to change the Python for build from python2.7 to the environment’s python3

    — Python 3:
    — Interpreter: /home/josefmtd/.virtualenvs/cv/bin/python3 (ver 3.6.5)
    — Libraries: /usr/lib/x86_64-linux-gnu/libpython3.6m.so (ver 3.6.5)
    — numpy: /home/josefmtd/.virtualenvs/cv/lib/python3.6/site-packages/numpy/core/include (ver 1.14.5)
    — packages path: lib/python3.6/site-packages
    —
    — Python (for build): /usr/bin/python2.7

    • Adrian Rosebrock

      The “Python (for build)” section is buggy and will not be correctly filled in. Ignore it. As long as your “Python 3” section is properly filled in you’ll be able to build the OpenCV bindings.

      • Hans

        Hi Adrian, here’s the thing: the Python 3 part is missing from the terminal output, which only shows Python (for build), even though I followed the instructions in the same order. Additionally, the cv2.cpython-36m-x86_64-linux-gnu.so is missing (in fact, the site-packages folder itself was missing.) Could those two errors be related, and if yes, what would be the way to fix them?

        • Adrian Rosebrock

          If the “Python 3” section is missing entirely then your .so bindings will not be built. Make sure you are in your Python virtual environment before running “cmake”.

    • Nathan Sieracki

      Josef,

      I encountered the same issue. Adrian mentioned this is of no consequence to the buid as long as the Python 3 section is filled in, but I resolved the Python (for build) issue by executing the below cmake tags, which pointed the python executable directory to the Python3, (note the final tag):

      cmake -D CMAKE_BUILD_TYPE=RELEASE \
      -D CMAKE_INSTALL_PREFIX=/usr/local \
      -D INSTALL_PYTHON_EXAMPLES=ON \
      -D INSTALL_C_EXAMPLES=OFF \
      -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.4.1/modules \
      -D PYTHON_EXECUTABLE=~/.virtualenvs/cv/bin/python3 \
      -D BUILD_EXAMPLES=ON
      -D BUILD_OPENCV_PYTHON3=YES

  18. jeremy

    Hi Adrian, i’ve followed your steps like a recipe book, but i have no idea why my Python (for build) is NO after finishing the cmake.

    i’ve tried other methods but it isn’t working.

    i am able to access python3.6.5 by typing ‘python’ in my terminal. 🙁

  20. CrashDummy

    Hi Adrian,

    Just setup Ubuntu 18.04 and your blog post email popped up in my inbox, amazing 🙂 !

    Sort of a side question, I’ve recently stumble across “pipenv” for setting up pip virtual environments, was wondering have you tried it and would love to hear your comments!

    • Adrian Rosebrock

      I’ve played with pipenv on one system but I personally haven’t put it through the ringer yet. I still really like virtualenv.

    • Adrian Rosebrock

      If you follow the instructions on this post you will have the C++ OpenCV installed as well. As far as your CMake errors go I cannot provide any suggestions without knowing what your error is.

  22. p pocock

    Thank You Guru Adrian! (guru means teacher)
    Your clear and straight-up tutorial worked for my Thinkpad P71, even after I installed ubunutstudio 18.4 next to Windows 10 Pro.

    No errors (a total rarity for me)!!

    Where do I find examples?

    Cheers Philip

    • Adrian Rosebrock

      Congrats on getting OpenCV installed on your system, that’s great! The PyImageSearch blog includes 250+ (free) tutorials on how to use computer vision and OpenCV in projects. If you are looking for a more structured, linear path to studying OpenCV I would recommend my book, Practical Python and OpenCV as well as the PyImageSearch Gurus course.

  23. Microsoft Support UK

    I don’t know how to configure and install the OpenCV in ubuntu. But after reading the post now I can install it in Ubuntu. I also need to install python for using the OpenCV.

    • Adrian Rosebrock

      The method detailed in this post will show you how to install Python and OpenCV on your Ubuntu system.

  24. clamytoe

    Hi Adrian,

    I’m using Anaconda with Linux Mint 19. and Just wanted to let others know that I was able to build OpenCV from source successfully by modifying the your instructions as follows:

    cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=~/anaconda3/envs/cv \
    -D INSTALL_PYTHON_EXAMPLES=ON \
    -D INSTALL_C_EXAMPLES=OFF \
    -D OPENCV_EXTRA_MODULES_PATH=~/Downloads/opencv_contrib-3.4.2/modules \
    -D PYTHON_EXECUTABLE=~/anaconda3/envs/cv/bin/python \
    -D BUILD_EXAMPLES=ON ..
    make -j4
    make install
    ldconfig -n ~/anaconda3/envs/cv/lib

    This of course assumes that Anaconda was installed in the default location. I put all of your instructions into a script, but I have not tested it.

    https://gist.github.com/clamytoe/3424d0201bba0073ff1313e80ffc6328

      • Tein

        Hi, in case I use the base environment of Anaconda (I don’t want to create a new environment), so can you tell me where the CMAKE_INSTALL_PREFIX and PYTHON_EXECUTABLE are?
        I assume that:
        -D CMAKE_INSTALL_PREFIX=~/anaconda3 \
        -D PYTHON_EXECUTABLE=~/anaconda3/bin/python \
        Is that right?
        Many tks

  25. Javier Hagua

    Hi again Adrian,
    Seems like it was some problem in the cmake part. I repeat the entire process and now it seems ok.
    Thanks for your private reply
    BR

  27. Alison

    hey how exactly are you supposed to edit the ~/.bashrc if workon command is not found? which part are you supposed to change?

    and if i chose not to use virtual, then what should i change?

    • Adrian Rosebrock

      You can use your favorite text editor to edit your .bashrc file. As far as changing it goes, without knowing what commands you’ve run on what errors you’ve encountered it’s hard for me to give any suggestions. Try to ensure your .bashrc file matches what I’ve included on the PyImageSearch blog. If you choose to not use virtual environments you can skip all ‘workon’ and ‘mkvirtualenv’ commands.

  28. Abhishek Singh

    In step 5, I am getting an error as unknown directory and i dont see any site-packages folder for python3.6
    Any help?

      • Adrian Rosebrock

        Hey Jacob, no need to submit three comments to the same post. I went ahead and removed the first, redundant one. If you have a “dist-packages” directory, which it sounds like you do, you can use that one instead.

  29. Anuj Pahade

    I’m getting this error

    mkvirtualenv cv -p python3
    ERROR: virtualenvwrapper could not find /usr/local/bin/virtualenv in your path

    which python
    /home/acer/anaconda3/bin/python

    Help needed

    • Adrian Rosebrock

      This tutorial was not designed for Anaconda. I would suggest using my exact install instructions.

  30. Jorge Paredes

    I compiled opencv 3.4.2 (I need it for your saliency-detection post 🙂 ), I follow your detailed instructions step by step and it works like a charm.

    Super blog!

    Thanks, Adrian.

  32. Hector Hernandez

    Great tutorial It saved me another painful nightmarish experience compiling opencv again….

    Just for the record, do NOT compile opencv in FAT32 systems, you might end up getting errors like this as the system is not able to set the executable bit on FAT 32 filesystems:

    Linking CXX shared library ../../lib/libopencv_core.so
    CMake Error: cmake_symlink_library: System Error: Operation not permitted
    CMake Error: cmake_symlink_library: System Error: Operation not permitted
    make[2]: *** [lib/libopencv_core.so.2.4.9] Error 1
    make[1]: *** [modules/core/CMakeFiles/opencv_core.dir/all] Error 2

    • Adrian Rosebrock

      Unfortunately you cannot. Python IDLE does not respect Python virtual environments. You should look into using Jupyter Notebooks instead.

  34. Andrea

    I had an issue at compile time with xfeatures2d. It threw a fatal error about a missing file “fatal error: opencv2/xfeatures2d/cuda.hpp: No such file or directory”.
    I solved it with a brute force approach: removing xfeature2d directory from contrib modules “rm -rf ~/opencv_contrib-3.4.1/modules/xfeatures2d/”.
    Maybe not the best way to solve the problem, but it compiles now.
    I suppose that xfeatures2d is needed for SIFT/SURF algorithms, so don’t remove this module unless you are not planning to use them!

  35. Eli

    Conflicts occur with Yolo v3. After googling it seems that it is better installing 3.4.0, but still not succeeding in compiling darknet yolo v3 with opencv. Anyone?

  36. Abdul

    Hello.
    Thanks for this great write up.
    I did follow the instruction without any problems.
    But, cv2.cpython-36m-x86_64-linux-gnu.so is not in /usr/local/lib/python3.6/site-packages/

    • Adrian Rosebrock

      Can you check and see if it’s in the build/lib directory? If so, you can manually copy it over.

  38. TACO_CODER

    Hello, sir thanks for your tutorial, what if i want to develop using CUDA? Do i need to do it on C++? can you please assist with the set up for CUDA please?

    Thanks in advance

  41. Jho

    First I had problem with not having /site-packages/ -folder, but I figured out I need to be installing opencv inside that virtualenv cv.(using ‘workon cv’) and then it worked fine.

    Second problem came, when I didn’t have ~/.virtualenvs/ -folder.
    It was located in root (/), not at home (~), which confused me when trying to link the cv2.so files. But linking to there it worked.

    • Adrian Rosebrock

      Hey Jho — it sounds like you may have accidentally ran all these steps in a root shell versus a user shell, that is likely what the discrepancy is.

  43. trico

    I tried this but its not working. Mainly:
    1) For the CMAKE options should we pass python libraries and python include dirs? Cmake complained about not finding them
    2) It found the python interpreter I gave but at the end I cannot find the .so file you mentioned. I do not have the site-packages folder anyway but rather a dist-packages one.
    3) As I cannot do the linking of the bindings then I cannot use open cv in the virtual environment.

    Thank you!

    • Adrian Rosebrock

      1. As long as you are in the “cv” Python virtual environment your Python libraries and include dirs should automatically be found cmake. You can pass them in manually but that should be unnecessary.

      2. You need the entire section filled out correctly, not just the interpreter. The include dirs, libraries, interpreter, and NumPy sections need to be filled out.

      3. See above.

  44. Jay

    Hi Adrian,
    I followed all the steps properly but I am now stuck at the build step. The error is,
    “CMake Error: The source directory “/home/jay/opencv-3.4.1″ does not appear to contain CMakeLists.txt.”
    I tried using the additional option -D for No such directory or file but it doesn’t seems to work as I am getting same error with or without additional -D option. If you can help it would be really great.

    • Adrian Rosebrock

      You’ll want to scroll up through your “cmake” output to find out exactly what the error is. The error you’re reporting in the comment is just a message from cmake saying the build files could not be created — the actual error is farther up in the output.

    • Bartolo

      There is a typo in the line “cmake -D CMAKE_BUILD_TYPE=RELEASE \”. It should be “cmake .. -D CMAKE_BUILD_TYPE=RELEASE \” because the previous command was “cd build”.

    • Tush

      Hi Jay,

      Looks like you forgot to do one step: $cd build

      your path should be like this: “/home/jay/opencv-3.4.1/build″

  45. Jay

    Hi Adrian,
    That’s what the problem is there is no output or anything just the error the build doesn’t even gets started. I even tried to start from scratch like setting up the vm again but still the error remains same.

    • Adrian Rosebrock

      Are you sure you’re in the “build” directory when you execute the “cmake” command? It almost sounds like your directory structure is incorrect. Delete the files and restart from where we download the OpenCV source code to our system.

  46. Rahat

    How can i add .virtualenvs on my pycharm ide?i mean how can run program terminal & pycharm in both side?

    • Adrian Rosebrock

      You’ll want to go to your project preferences and set then the “Python Interpreter” to point to whatever Python virtual environment you are using.

  48. Patrick

    Hello Adrian

    Are you aware that we need to add this switch in CMAKE now to get the “non-free” modules ?

    -D OPENCV_ENABLE_NONFREE:BOOL=ON \

    Best Regards

  49. Byron

    Hi Adrian,
    When compiling OpenCV with ‘make’ I run into the error “unsupported GNU version! gcc versions later than 6 are not supported!”. I’ve tried reading online where I could perhaps use an earlier GNU version with no success. Could you please assist?
    Regards
    Byron

    • Adrian Rosebrock

      I assume you’re running GCC-7? If so, you can install GCC-6 via:

      $ sudo apt-get install gcc-6 g++-6

    • Eli

      When you want to use a different than default gcc and g++ you want to set the environment variables CC and CXX which cmake uses to determine which compiler to use. For example “export CXX=/usr/bin/g++-6”

      • Eli

        I believe the standard ‘make’ from the default config uses the CC and CXX as well… Also for the CC variable for example it would be “export CC=/usr/bin/gcc-6”

  50. Rahul

    Hi Adrian ,

    I read this tutorial and thanks for sharing. I had actually started using 18.04 and have also configured with cuda 10.0 along cudNN_7.4 . Now I noticed about your suggestion to stick on with 16.04 for deep learning purposes. Will all the codes which are working with ubuntu 16.04 and previous versions of cuda work fine with this new setup ? Or do you suggest me to shift back to 16.04 ? (PS:I actually do not have any particular reason for using ubuntu 18.04)

    Thanks
    Rahul

    • Adrian Rosebrock

      I’ll actually be publishing a new tutorial on Ubuntu 18.04 + CUDA for deep learning in the next couple of weeks. The steps to install are similar but there are a few things to watch out for.

  51. Rob

    Hi Adrian,
    Great Tutorial.
    However, I’m getting the following error when attempting to run one of your examples..

    …

    (Image:78): Gtk-WARNING **: 23:47:14.812: cannot open display:

    • Adrian Rosebrock

      Are you trying to run the script over SSH? If so, you need to enable X11 forwarding:

      $ ssh -X user@your_ip_address

  52. Aayush Kubba

    I am unable to find cv2.cpython-36m-x86_64-linux-gnu.so in any of the directory. I am using Linux Mint.

  53. Ben Gosney

    tl;dr
    sudo apt install libatlas-base-dev libjasper-dev libqtgui4 python3-pyqt5 libqt4-test && sudo pip3 install opencv-python
    takes a few mins to run, and is STABLE

    • Adrian Rosebrock

      Sorry, that’s not entirely true and will lead you to trouble. If you would like install OpenCV via pip you should see this guide instead. Keep in mind that you are not installing any of the “contrib” packages which you should do if you plan on OpenCV. But, even if you do perform a pip install of OpenCV you won’t have access to the NONFREE modules which this blog post demonstrates how to do. Keep in mind that the two are not the same.

  54. Mehul Bhardwaj

    sudo pip install virtualenv virtualenvwrapper

    In this command, when I simply used pip I got an error in

    source ~/.bashrc

    Then I used pip3 and didn’t get the error. Is that something you need to update or was it something that I did wrong?

    • Adrian Rosebrock

      Double-check that your .bashrc file updates match mine inside the post. Provided you can run “source ~/.bashrc” without an error you should be okay.

  55. Brian Tremaine

    Hi Adrian,

    Thank you so much for this great blog. This saved me!!
    Before I saw this I had installed Ubuntu 18.04 under a Windows subsystem for Linux on my Windows 10 system along with the virtualenv and virtualenvwrapper. I was stuck on installing OpenCV and tried several pip and get-apt approaches. Your blog saved me. I followed the steps for the CV2 build and install and it seems to all be working now.

    • Adrian Rosebrock

      Thanks for the kind words, Brian! And congratulations on getting OpenCV installed on your Ubuntu system!

  56. Afedisa

    Great job!!
    I have a problem, I have Ubuntu 18 running on an odroid and when I try to install opencv but it’s not possible because it can’t find jasper libs for this Ubuntu.
    Any idea?

  57. Maryam

    Hello Adrian,
    thanks for the useful tutorial. I am a novice in work with image and video but it is not a convinced reason to my proffesor for not handing in the project on time.
    I have used this app face-recognition ” https://github.com/ageitgey/face_recognitionhttps://github.com/ageitgey/face_recognition“.
    I applied ubuntu 18.04 and use camera of laptop to implement this project. I have already installed dilib, cmake and open cv to detect and recognized faces on a realtime video.
    unfortunately I faced two problems that I have no idea about how I can solve it.
    the problems are as follows:
    I want to write a code in order to write the time and date which camera recognize the face of that person.
    I wanna create a database or a folder which contains face of the recognized person with the time and date.

    I know that you are busy but I think you are so generous to spend your time to give me a tutorial which contains the code that has the ability to show the time and date.
    I am really looking forward to your answer.
    Best Regards
    Maryam

  58. Muflih

    Hi Adrian!

    Thanks for helpful tutorial, it works well for me

    But it is possible to remove virtualenv after installation?

    Thanks..

    • Adrian Rosebrock

      Technically yes, but I don’t recommend that as Python virtual environments are a best practice. That said — once the “cv2.so” bindings are in your system “site-packages” directory you could technically remove the virtual environment (NOT recommended).

  59. Yash Jain

    Hi Adrian
    Thanks for , such a great tutorial for installing opencv3.4.4 on Ubutnu 18.04.
    I installed Opencv successfully on aws ubuntu18.04 . with make
    followed every step as you suggest.
    I am accessing my aws ubuntu server using ssh on window 10 (laptop dell having front camera) and trying to access camera using VideoCapture(-1) function.but getting error (VIDEOIO ERROR: V4L: can’t find camera device).

    Please suggest what i am missing .
    i also enable x11 forwarding in ssh.
    installed xming.

  60. Björn

    Following your steps, when I get to the CMake part, niether python2.7 nor python3.5 show up. I have no idea what im doing wrong, any help would be greatly appriciated!

    If its of importance I am using Raspian Lite.

    • Adrian Rosebrock

      Make sure you are inside the Python virtual environment before executing “cmake”.

    • Adrian Rosebrock

      Python IDLE does not respect virtual environments. If you really want an IDLE-like environment you should use Jupyter Notebooks.

  62. Shankar

    For OpenCV 4.0.1 I am comping now, on Linux, looks like the default support for pkg-config is gone. You need to explicitly ask for it. If you like that then use the following cmake command:

    cmake -D CMAKE_BUILD_TYPE=RELEASE \
	-D CMAKE_INSTALL_PREFIX=/opt/opencv \
	-D INSTALL_PYTHON_EXAMPLES=ON \
	-D INSTALL_C_EXAMPLES=OFF \
        -D OPENCV_ENABLE_NONFREE=ON \
        -D OPENCV_GENERATE_PKGCONFIG=ON \
	-D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-4.0.1/modules \
	-D PYTHON_EXECUTABLE=/opt/anaconda3/envs/OpenCV/bin/python \
	-D BUILD_EXAMPLES=ON ..

    opencv4.pc file should be generated in …/opencv/lib/pkgconfig/ directory.

  63. Raj

    Hi Adrian, the guide was thorough but i am facing issues in ‘cmake’ process. After executing the command as it is, the output was inconclusive and did not ended as it should according to guide. Please try and solve my problem as soon as possible.
    Note: Newbie here.

  64. Klifford

    Hi Adrian, the guide is well done! However, I am facing a symlink problem. When I try to import cv2, it says cannot open shared object file: no such file or directory. I checked that I have the original file and the symlink. I think I missed a step or so. Any help would be appreciated

    • Adrian Rosebrock

      That is indeed a sym-link issue. Double-check the sym-link path. You can delete the sym-link and recreate it as well.

  65. Chim

    Hello Adrian, thank you for this great tutorial.
    Once I created the cv env, how can I chose it in pycharm as env? I’m struggling with this.
    Thank you!

  66. Johnson Ho

    Thank you! This worked perfectly as far as I can tell. Can’t wait to move forward, following some of your other guides!

  67. Paul

    Hi Adrian,

    Thanks a lot for the tutorial! I installed it on Ubuntu 18.04 and I had some trouble to build with cmake. I got an error at about 10% because of Eigen packet.

    I managed to disable it (WITH_EIGEN=OFF) thans to ccmake (see https://cmake.org/runningcmake/). Now everything is working fine!

    Maybe it can help someone facing this issue.

    Best regards,

    Paul

  68. akash

    so now if i want to code python and import cv2 i need to code on terminal and not on a user friendly IDE?

    • Adrian Rosebrock

      I’m not sure I understand. The method covered in this post will work when importing via the terminal.

      • pearl

        hi adrian,
        can you please tell me how to safely uninstall opencv 3.4.4 and reinstall again

        Also i could not find libopencv_core, libopencv_highgui and libopencv_imgproc in the installed opencv 3.4.4. Can you please give me a solution. How should i cmake

        • Adrian Rosebrock

          Delete your “build” directory, re-create it, re-run “cmake” and “make”. Then re-run “sudo make install” to overwrite the OpenCV install with your new install.

  70. Vinit

    Hi Dr. Adrian,

    I recently enrolled in the 17 day crash course of yours. My question for you is, can setup the same environment on google colab? if yes then how? if not then where else? something similar to google colab, and something that will not cost a lot of money like AWS or Azure.

  71. Nathan

    Hi Adrian,

    Just wanted to thank you for being so short and yet precise in your steps. Everything worked fine straight off the bat and it’s great to have people like you explain stuff like this because you don’t come across as condescending or give convoluted explanations, just short, sweet and precise.

    So thanks again! 😀

  72. nquantum

    I curous why you install virtualenv, virtualenvwrapper, numpy from pip (python2) than pip3 (python3)?

    It should not benefit by install from pip3?

    • Adrian Rosebrock

      In Ubuntu 18.04 the default Python install is Python 3 therefore “pip” also points to Python 3.

  73. bluenotes

    Great tutorial!
    Is there a way to say compile OpenCV for Python on just once on one computer but installing it on muliple computer (i.e. just copy specific files and then install in virtual environment)?
    Many Thanks

    • Adrian Rosebrock

      This method will install OpenCV for C/C++ as well. A Python virtual environment is not needed for C/C++ development.

  76. Heinz Ambach

    Hi Adrian,

    I’m a newbie to python and opencv. I installed opencv as per your instructions and can access it via makeon at the linux prompt. I like using anaconda-navigator, but cannot fathom how to access opencv (import cv?) within spyder. When I check the environments in anaconda-navigator, looking for opencv. I find version 3.4.2 whereas the above procedure installed 3.4.4.
    Is there a quick solution?
    Thanks in advance

  77. adnan

    Hi Adrian,
    Thanks a lot for this amazing tutorial, Really loved it!!

    Except there is one problem I’m not been able to figure out.

    I’ve gone through each and every step as you have instructed but after running the CMake, the output at the terminal shows no block for python3 at all, Instead, everything that should have been under the python3 is displayed under python2.7.

    Further after compiling and linking with the virtual machine, the import command

    “import cv2” works only on python2.7 and not on any version of python3 .

    I’ve gone through all the steps once again to make any possible revisions but have failed to find any possible solution to this problem on my own.

    please help me with this

    • Adrian Rosebrock

      It seems like that you accidentally created a Python virtual environment for Python 2.7 rather than Python. Make sure you follow the “mkvirtualenv” step carefully and ensure you specific the correct Python version.

Comment section

Hey, Adrian Rosebrock here, author and creator of PyImageSearch. While I love hearing from readers, a couple years ago I made the tough decision to no longer offer 1:1 help over blog post comments.

At the time I was receiving 200+ emails per day and another 100+ blog post comments. I simply did not have the time to moderate and respond to them all, and the sheer volume of requests was taking a toll on me.

Instead, my goal is to do the most good for the computer vision, deep learning, and OpenCV community at large by focusing my time on authoring high-quality blog posts, tutorials, and books/courses.

If you need help learning computer vision and deep learning, I suggest you refer to my full catalog of books and courses — they have helped tens of thousands of developers, students, and researchers just like yourself learn Computer Vision, Deep Learning, and OpenCV.

Click here to browse my full catalog.

Similar articles

You can learn Computer Vision, Deep Learning, and OpenCV.

Get your FREE 17 page Computer Vision, OpenCV, and Deep Learning Resource Guide PDF. Inside you’ll find our hand-picked tutorials, books, courses, and libraries to help you master CV and DL.

Already a member of PyImageSearch University? Click here to login.

Access the code to this tutorial and all other 500+ tutorials on PyImageSearch

Enter your email address below to learn more about PyImageSearch University (including how you can download the source code to this post):

What's included in PyImageSearch University?

  • Easy access to the code, datasets, and pre-trained models for all 500+ tutorials on the PyImageSearch blog
  • High-quality, well documented source code with line-by-line explanations (ensuring you know exactly what the code is doing)
  • Jupyter Notebooks that are pre-configured to run in Google Colab with a single click
  • Run all code examples in your web browser — no dev environment configuration required!
  • Support for all major operating systems (Windows, macOS, Linux, and Raspbian)
  • Full access to PyImageSearch University courses
  • Detailed video tutorials for every lesson
  • Certificates of Completion for all courses
  • New courses added every month! — stay on top of state-of-the-art trends in computer vision and deep learning

PyImageSearch University is really the best Computer Visions "Masters" Degree that I wish I had when starting out. Being able to access all of Adrian's tutorials in a single indexed page and being able to start playing around with the code without going through the nightmare of setting up everything is just amazing. 10/10 would recommend.

Sanyam BhutaniMachine Learning Engineer and 2x Kaggle Master
Close