[python] How do I install Python OpenCV through Conda?

I'm trying to install OpenCV for Python through Anaconda, but I can't seem to figure this out.

I tried

conda install opencv
conda install cv2

I also tried searching

conda search cv

No cigar. I ran across this which lists opencv as an included package:

http://docs.continuum.io/anaconda/pkgs.html

After running conda info I noticed my version is 3.4.1, but I couldn't seem to find any information about this version online. I'm pretty confused about this.

Am I missing something pretty obvious here? If opencv was available for a previous version of Anaconda, then why wouldn't it be available for the newer version? And why does that link only show me documentation for version 1.9.2?

I have just tried on two Win32 Python 3.5 computers. At first, I was able to conda install opencv, but it didn't work nor did the version from menpp, but this did:

conda install -c https://conda.binstar.org/conda-forge opencv

i was on MAC machine inside one of anaconda virutal environment. For me,

conda install -c conda-forge opencv

worked fine.

It installed opencv version 3.4.4

Hope it helps.

I tried to use Minoconda on my Raspberry Pi (Raspibian OS), but I could not get Open CV for my platform. Finally the following command worked:

$ sudo apt-get install ipython python-opencv python-scipy python-numpy python-setuptools python-pip

It doesn't seem like the page you linked includes opencv any more. (Funny, I do recall it being included at a previous point as well.)

In any case, installation of OpenCV into Anaconda, although unavailable through conda, is pretty trivial. You just need to download one file.

1. Download OpenCV from https://opencv.org/releases/ and extract
2. From the extracted folder, copy the file from the extracted directory: opencv/build/python/2.7/(either x86 or x64, depending on your Anaconda version)/cv2.pyd to your Anaconda site-packages directory, e.g., C:\Anaconda\Lib\site-packages
3. To get FFmpeg within opencv to work, you'll have to add the directory that FFmpeg is located in to the path (e.g., opencv/sources/3rdparty/ffmpeg). Then you'll have to find the DLL file in that folder (e.g., opencv_ffmpeg_64.dll) and copy or rename it to a filename that includes the opencv version you are installing, (e.g., opencv_ffmpeg249_64) for 2.4.9.

Now at the Python prompt you should be able to type "import cv2"...to verify that it works, type "print cv2.__version__", and it should print out the OpenCV version you downloaded.

Although not through Conda, and this is specific to Ubuntu, the easiest way to install OpenCV with all its contrib modules built in such as SIFT and SURF is to use:

pip install opencv-contrib-python

It supports Python2.7+ and Python 3.4+

If you choose not to have the contribs built in, you could instead do:

pip install opencv-python

You need to install the following dependencies below before you run those commands though:

sudo apt-get install build-essential cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev

This is a good tutorial on setting OpenCV on Ubuntu: https://medium.com/@debugvn/installing-opencv-3-3-0-on-ubuntu-16-04-lts-7db376f93961

The following command works for me too. I am using an embedded IPython Notebook in Anaconda.

conda install -c https://conda.binstar.org/menpo opencv

conda install -c https://conda.anaconda.org/menpo opencv

Try this

To install the OpenCV package with conda, run:

conda install -c menpo opencv3=3.1.0

https://anaconda.org/menpo/opencv3

On Linux, as discussed here, the best way to get opencv at present is from loopbio at conda-forge:

conda install -c loopbio -c conda-forge -c pkgw-forge ffmpeg-feature ffmpeg gtk2 opencv

If you have 'a modern CPU' there exists also a compiled version "enabling all modern CPU instruction set extensions [...] and against libjpeg-turbo":

conda install -c loopbio -c conda-forge -c pkgw-forge ffmpeg-feature ffmpeg gtk2 opencv-turbo`

Two of the solutions mentioned in other answers don't work unconditionally:

• The conda you get through conda install opencv or pip install opencv-python doesn't have gtk2 support, so you can't display images through imshow.
• Conda built by Menpo (conda install -c menpo opencv3) has gtk2 support, but
  • they have only built OpenCV 3.2 for Python 3.5, not Python 3.6
  • Ubuntu 16.10 has deprecated libpng12, leading to a missing dependency and the following error when trying to import cv2: ImportError: libpng12.so.0: cannot open shared object file: No such file or directory, as discussed here

EDIT: @Yamaneko points out that as of 14 June 2017, "there is an issue with loopbio's OpenCV version 3.2.0. It silently fails to read and write videos. It is due to a combination of an upstream OpenCV issue in combination with an old GCC. More details here. There is a PR on its way to solve the issue."

To install OpenCV in Anaconda, start up the Anaconda command prompt and install OpenCV with

conda install -c https://conda.anaconda.org/menpo opencv3

Test that it works in your Anaconda Spyder or IPython console with

import cv2

You can also check the installed version using:

cv2.__version__

I faced the same problem, but I have solved it now. This is what I did:

First enter conda install -c https://conda.binstar.org/menpo opencv in the command prompt and then find the path Anaconda\pkgs\opencv-2.4.9.1-np19py27_0\Lib\site-packages. Now copy all the files present here into Anaconda\Lib\site-packages. Now you will be able to use OpenCV with Python.

If you want to install opencv 3.4.0, there, unfortunately, is not this version inside conda. You need to use pip instead.

pip install opencv-python==3.4.0.12

The correct command for installing the current version of OpenCV 3.3 in Anaconda windows:

conda install -c conda-forge opencv

or

conda install -c conda-forge/label/broken opencv

For 3.2 use this:

conda install -c menpo opencv3

If conda install opencv or conda install -c https://conda.binstar.org/menpo opencv does not work, you can try to compile from the source.

Download the source from http://opencv.org/downloads.html, follow the install instruction in http://docs.opencv.org/2.4/doc/tutorials/introduction/linux_install/linux_install.html, (maybe you can jump to the last part directly, 'Building OpenCV from Source Using CMake...), change the cmake command as following:

mkdir release
cd release
cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/home/**/env/opencv-2.4.10 -D BUILD_NEW_PYTHON_SUPPORT=ON -D PYTHON_EXECUTABLE=/home/**/env/anaconda/bin/python -D PYTHON_INCLUDE_DIR=/home/**/env/anaconda/include/python2.7 -D PYTHON_LIBRARY=/home/**/env/anaconda/lib/libpython2.7.so -D PYTHON_PACKAGES_PATH=/home/**/env/anaconda/lib/python2.7/site-packages -D PYTHON_NUMPY_INCLUDE_DIRS=/home/**/env/anaconda/lib/python2.7/site-packages/numpy/core/include ..

make -j4
make install

You will find cv2.so in anaconda/lib/python2.7/site-packages.

Then:

import cv2
print cv2.__version__

It will print out 2.4.10.

My environment is GCC 4.4.6, Python 2.7 (anaconda), and opencv-2.4.10.

I think the easiest option that should work cross-platform and across various versions of Anaconda is

#Run Anaconda Prompt as an Administrator (on Windows) & execute the command below
#(you may need to use   sudo   on a Unix-Based or Mac system 
#to have the needed permissions)

conda install -c conda-forge opencv

Like others, I had issues with Python 3.5.1/Anaconda 2.4.0 on OS X 10.11..

But I found a compatible package here:

https://anaconda.org/menpo/opencv3

It can be installed via the command line like so:

 conda install -c https://conda.anaconda.org/menpo opencv3

Worked like a charm. First time I've ever gotten OpenCV to work on 3.x!

You can install OpenCV by running these commands in the Anaconda command prompt:

conda config --add channels conda-forge

conda install libopencv opencv py-opencv

Source:

https://github.com/conda-forge/opencv-feedstock

Here's a general approach to using conda to install packages for Python that applies:

1. conda search packageName e.g. conda search opencv

2. If this doesn't return results, conda install packageName will not work.

3. At this point you can go to, https://anaconda.org/ and type the packageName into the search box. If this pulls up results (which it should for OpenCV), then click on one of the results that is for your platform (e.g. win-64). The next page will show you the command to use to install this package (e.g. conda install -c menpo opencv=2.4.11).

4. If your package doesn't return results by search https://anaconda.org, then you can try pip install packageName.

Caution: when I used step 3 to install OpenCV for win-64, I got an error when I tried to import cv2.

Here is the error:

RuntimeError: module compiled against API version a but this version of numpy is 9
Traceback (most recent call last):
File "<stdin>", line 1, in <module>

ImportError: numpy.core.multiarray failed to import

I think the error is due to package version conflicts. Nevertheless, this is a valid way to install OpenCV and other Python packages, you just might need to resolve some package version conflicts.

You just copy the cv2.pyd file to the C:\Users\USERNAME\Anaconda2\Lib directory.

You get the cv2.pyd file at this link (https://sourceforge.net/projects/opencvlibrary/files/).

The cv2.pyd is located at C:\Users\USERNAME\Desktop\opencv\build\python\2.7\x64.

Following command adds a different anaconda channel for opencv3, you should be able to pull from it.

conda install --channel  https://mirrors.ustc.edu.cn/anaconda/cloud/menpo opencv3

I had exactly the same problem, and could not get conda to install OpenCV. However, I managed to install it with the OpenCV installer you find at this site:

http://www.lfd.uci.edu/~gohlke/pythonlibs/

His files are "Wheel" whl files that can be installed with pip, e.g.

pip install SomePackage-1.0-py2.py3-none-any.whl

in a command window. It worked with Spyder directly after executing this command for me. I have had the same experience with other packages, the above UC Irvine site is a gold mine.

Windows 7, conda 4.4.10 python 2.7.6 Downloaded opencv package from Unofficial Windows Binaries for python extensions packages. (picked python 2.4 AMD 64) cd Download pip install opencv_python... python run_some.py (where import cv2 ...) seems to work. YMMV

conda install opencv currently works for me on UNIX/python2. This is worth trying first before consulting other solutions.

In May 2019, the answer is as follows for MacOS:

conda update freetype
conda install opencv -c conda-forge

This updates libfreetype.6.dylib to 24.0.0 compatibility and installs OpenCV 4.1.0.

it works on anaconda3 windows 10 I have already downloaded it at 5 December 2019.

Firstly, using this command:

pip install opencv-contrib-python

after that windows will ask for permission and try again:

pip install opencv-contrib-python --user

look at this it works!!

I tried following command and it works fine

conda install -c conda-forge opencv

once you hit the command it will ask for yes or no

If you select yes it will proceed and install all the required packages

You can install it using binstar:

conda install -c menpo opencv

I have summarized my now fully working solution, OpenCV-Python - How to install OpenCV-Python package to Anaconda (Windows). Nevertheless I've copied and pasted the important bits to this post.

At the time of writing I was using Windows 8.1, 64-bit machine, Anaconda/ Python 2.x. (see notes below - this works also for Windows 10, and likely Python 3.x too).

• NOTE 1: as mentioned mentioned by @great_raisin (thank you) in comment section however, this solution appears to also work for Windows 10.

• NOTE 2: this will probably work for Anaconda/Python 3.x too. If you are using Windows 10 and Anaconda/Python 3.x, and this solution works, please add a comment below. Thanks! (Update: noting from comment "Working on Windows 10")

• NOTE 3: depending on whether you are using Python 2.x or 3.x, just adjust the print statement accordingly in code snippets. i.e. in Python 3.x it would be print("hello"), and in Python 2.x it would be print "hello".

TL;DR

To use OpenCV fully with Anaconda (and Spyder IDE), we need to:

1. Download the OpenCV package from the official OpenCV site
2. Copy and paste the cv2.pyd to the Anaconda site-packages directory.
3. Set user environmental variables so that Anaconda knows where to find the FFMPEG utility.
4. Do some testing to confirm OpenCV and FFMPEG are now working.

(Read on for the detail instructions...)

Prerequisite

Install Anaconda

Anaconda is essentially a nicely packaged Python IDE that is shipped with tons of useful packages, such as NumPy, Pandas, IPython Notebook, etc. It seems to be recommended everywhere in the scientific community. Check out Anaconda to get it installed.

Install OpenCV-Python to Anaconda

Cautious Note: I originally tried out installing the binstar.org OpenCV package, as suggested. That method however does not include the FFMPEG codec - i.e. you may be able to use OpenCV, but you won't be able to process videos.

The following instruction works for me is inspired by this OpenCV YouTube video. So far I have got it working on both my desktop and laptop, both 64-bit machines and Windows 8.1.

Download OpenCV Package

Firstly, go to the official OpenCV site to download the complete OpenCV package. Pick a version you like (2.x or 3.x). I am on Python 2.x and OpenCV 3.x - mainly because this is how the OpenCV-Python Tutorials are setup/based on.

In my case, I've extracted the package (essentially a folder) straight to my C drive (C:\opencv).

Copy and Paste the cv2.pyd file

The Anaconda Site-packages directory (e.g. C:\Users\Johnny\Anaconda\Lib\site-packages in my case) contains the Python packages that you may import. Our goal is to copy and paste the cv2.pyd file to this directory (so that we can use the import cv2 in our Python codes.).

To do this, copy the cv2.pyd file...

From this OpenCV directory (the beginning part might be slightly different on your machine). For Python 3.x, I guess, just change the 2.x to 3.x accordingly.

# Python 2.7 and 32-bit machine:
C:\opencv\build\python\2.7\x84

# Python 2.7 and 64-bit machine:
C:\opencv\build\python\2.7\x64

To this Anaconda directory (the beginning part might be slightly different on your machine):

C:\Users\Johnny\Anaconda\Lib\site-packages

After performing this step we shall now be able to use import cv2 in Python code. BUT, we still need to do a little bit more work to get FFMPEG (video codec) to work (to enable us to do things like processing videos).

Set Environmental Variables

Right-click on "My Computer" (or "This PC" on Windows 8.1) ? left-click Properties ? left-click "Advanced" tab ? left-click "Environment Variables..." button.

Add a new User Variable to point to the OpenCV (either x86 for 32-bit system or x64 for 64-bit system). I am currently on a 64-bit machine.

| 32-bit or 64 bit machine? | Variable     | Value                                |
|---------------------------|--------------|--------------------------------------|
| 32-bit                    | `OPENCV_DIR` | `C:\opencv\build\x86\vc12`           |
| 64-bit                    | `OPENCV_DIR` | `C:\opencv\build\x64\vc12`           |

Append %OPENCV_DIR%\bin to the User Variable PATH.

For example, my PATH user variable looks like this...

Before:

C:\Users\Johnny\Anaconda;C:\Users\Johnny\Anaconda\Scripts

After:

C:\Users\Johnny\Anaconda;C:\Users\Johnny\Anaconda\Scripts;%OPENCV_DIR%\bin

This is it we are done! FFMPEG is ready to be used!

Test to confirm

We need to test whether we can now do these in Anaconda (via Spyder IDE):

• Import OpenCV package
• Use the FFMPEG utility (to read/write/process videos)

Test 1: Can we import OpenCV?

To confirm that Anaconda is now able to import the OpenCV-Python package (namely, cv2), issue these in the IPython console:

import cv2
print cv2.__version__

If the package cv2 is imported OK with no errors, and the cv2 version is printed out, then we are all good! Here is a snapshot:

_{(source: mathalope.co.uk)}

Test 2: Can we Use the FFMPEG codec?

Place a sample input_video.mp4 video file in a directory. We want to test whether we can:

• read this .mp4 video file, and
• write out a new video file (can be .avi or .mp4 etc.)

To do this we need to have a test Python code, call it test.py. Place it in the same directory as the sample input_video.mp4 file.

This is what test.py may look like (I've listed out both newer and older version codes here - do let us know which one works / not work for you!).

(Newer version...)

import cv2
cap = cv2.VideoCapture("input_video.mp4")
print cap.isOpened()   # True = read video successfully. False - fail to read video.

fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter("output_video.avi", fourcc, 20.0, (640, 360))
print out.isOpened()  # True = write out video successfully. False - fail to write out video.

cap.release()
out.release()

(Or the older version...)

import cv2
cv2.VideoCapture("input_video.mp4")
print cv2.isOpened()   # True = read video successfully. False - fail to read video.

fourcc = cv2.cv.CV_FOURCC(*'XVID')
out = cv2.VideoWriter("output_video.avi",fourcc, 20.0, (640,360))
print out.isOpened()  # True = write out video successfully. False - fail to write out video.

cap.release()
out.release()

This test is VERY IMPORTANT. If you'd like to process video files, you'd need to ensure that Anaconda / Spyder IDE can use the FFMPEG (video codec). It took me days to have got it working. But I hope it would take you much less time! :)

Note: One more very important tip when using the Anaconda Spyder IDE. Make sure you check the current working directory (CWD)!!!

Conclusion

To use OpenCV fully with Anaconda (and Spyder IDE), we need to:

1. Download the OpenCV package from the official OpenCV site
2. Copy and paste the cv2.pyd to the Anaconda site-packages directory.
3. Set user environmental variables so that Anaconda knows where to find the FFMPEG utility.
4. Do some testing to confirm OpenCV and FFMPEG are now working.

Good luck!

Using Wheel files is an easier approach. If you cannot install Wheel files from the command prompt, you can use an executable pip file which exists in the <Anaconda path>/Scripts folder.

I had steps in repo that will help you properly install OpenCV.

If the links if broken, see this.

For requirements and prerequisites follow this instructions:

YOU NEED TO HAVE WINDOWS OS to begin this project.

windows 7 to 10

Follow the steps carefully

1. Install Anaconda2 32/64 python 2

   or

   Install Anaconda3 32/64 python 3

   It Depends on System Information

   Along the way you must check all the options ignore the red text

   Then wait till installation completed ...

   Open CMD admin when finish installed

   Make sure that the anaconda-python is configured in CMD admin

2. Type: python

   This will come out:

   C:\WINDOWS\system32>python Python 3.6.3 |Anaconda custom (64-bit)| (default, Oct 15 2017, 03:27:45)

   [MSC v.1900 64 bit (AMD64)] on win32 Type "help", "copyright", "credits" or "license" for more information.

3. Next is to install packages:

   Numpy

   OpenCV

   PyMySQL or pymysql

   Install and Import OpenCV

   conda install -c conda-forge opencv = (version optional)

   (click yes if there's updates)

   Install and Import Numpy

   pip install numpy

   (click yes if there's updates)

   Install and Import PyMySQL/pymysql

   pip install PyMySQL or pymysql

4. When all the package is completely installed, then you must Reboot it.

   go to CMD admin again

   type "python"

   import cv2 (cv2 is OpenCV)

   import Numpy or numpy

   import pymysql

   import os

   from PIL from Image

   if all of then is ok.. your ready to run the code!!

   After you finish all the steps, you can now test the code in SPYDER python that I use

I just installed conda 4.7.11 in Windows 10. OpenCV can be easily installed in Anaconda Navigator. After launching Navigator, click on Environments on the left panel. In the top drop-down, select Not installed. Then search for py-opencv, tick on the left checkbox to install it. It will install the dependent package libopencv. I use import cv2 in Spyder to access it.

One thing i think nobody mentioned is that if you are using Windows N or KN edition, please install Media Feature Pack that is used for OpenCV. Also make sure you have Visual C++ Distributable installed. And then you can use command

conda install -c menpo opencv

Windows only solution. OpenCV 3.x pip install for Python 3.x

Download .whl file (cpMN where you have Python M.N). contrib includes OpenCV-extra packages. For example, assuming you have Python 3.6 and Windows 64-bit, you might download opencv_python-3.2.0+contrib-cp36-cp36m-win_amd64.whl

From command prompt type:

pip install opencv_python-3.2.0+contrib-cp36-cp36m-win_amd64.whl

You'll have a package in your conda list : opencv-python 3.2.0+contrib <pip>

Now you could test it (no errors):

>>> import cv2
>>>

Original source page where I took the information is here.

I entered the following command in the command prompt:

conda install -c menpo opencv=2.4.11

This worked for me!!!

An easy and straight forward solution is to install python3.5 first before installing opencv3

conda install python=3.5

followed by

conda install --channel https://conda.anaconda.org/menpo opencv3 

I installed it like this:

$ conda install --channel https://conda.anaconda.org/conda-forge opencv

I tried conda install opencv directly, but it does not work for me since I am using Python 3.5 which is higher version that default OpenCV library in conda. Later, I tried 'anaconda/opencv', but it does not work either. I found finally that conda-forge/opencv works for Python 3.5.

I just wanted to update the brilliant answer by Atlas7.

If you're using OpenCV 3, change the test code to the following:

import cv2
cap=cv2.VideoCapture("input_video.mp4")
print cap.isOpened()   # True = read video successfully. False - fail to read video.

fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter('output.avi',fourcc, 20.0, (640,480))
print out.isOpened()  # True = write out video successfully. False - fail to write out video.

cap.release()
out.release()

This worked for me (on Ubuntu and conda 3.18.3):

conda install --channel https://conda.anaconda.org/menpo opencv3

The command above was what was shown to me when I ran the following:

anaconda show menpo/opencv3

This was the output:

To install this package with conda run:
     conda install --channel https://conda.anaconda.org/menpo opencv3

I tested the following in python without errors:

>>> import cv2
>>>

Try following these commands in this order:-

conda update anaconda-navigator
conda update navigator-updater
pip install opencv-python

The following installs opencv from conda-forge (note: tried on Windows)

conda config --add channels conda-forge
conda install opencv

To install OpenCv with conda on Windows-64 and Python 3.5, the only one that worked for me is:

conda install -c conda-forge opencv=3.1.0