Install Python

We have 2 main versions of Python: Python 2.x and Python 3.x. 3.x is more popular. (see the difference)

Install Python directly:

Note: Choose stable releases and compatible to your OS (Windows, MAC or Linux), recommend for “executable installer”.

  • Download and install it, check in “Add to PATH” box to make sure that you can work more convenience through Command Prompt window.

  • After finishing installation. Let’ check it by accessing Command Prompt window and type python. when the prompt of the interactive Python interpreter ”»>” appears, it is successful.

C:\Users\admin>python
Python 3.6.8 (tags/v3.6.8:3c6b436a57, Dec 24 2018, 00:16:47) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>>

Here, we can use Python IDLE (Integrated DeveLopment Environment) as an editor.

Install OpenCV

  • Install OpenCV library: First, we have to install preliminary library. 
    pip install numpy
pip install matplotlib

    Note: PIP is a package manager for Python packages, or modules.

Download and install through this link to get .whl file, this is wheel format as a standard build-package format for Python. At this site, many standard libraries are present.

  • To check it, access Python type 2 command” 
    >>> import cv2
>>> cv2.__version__

Install Python using Anaconda:

[Anaconda] is an open-source distribution of Python and R. It aims to simplify package management and deployment.

Python conda can be installed in the process which sets up Anaconda.

  • Access Anaconda.com, choose Downloads label and consider your platform and Python version.

  • Run installation file and customize options. Add Anaconda to the system PATH environment for using directly on command prompt window. Otherwise, anaconda prompt is also useful.

  • To check the installation, type command conda info in command prompt window:

  • Next, we create environment to manage package with command conda create --name <your_env_name>. This way will generate an environment and contain package library.

## Package Plan ##

  environment location: D:\Anaconda3\envs\new-env

Proceed ([y]/n)? y

Preparing transaction: done
Verifying transaction: done
Executing transaction: done
#
# To activate this environment, use
#
#     $ conda activate new-env
#
# To deactivate an active environment, use
#
#     $ conda deactivate

conda activate new-env to access the environment and conda deactivate new-env to quit out.

  • Install OpenCV by command conda install -c conda-forge opencv. This solution is easier than install as a global library. However, access in environment is required in every single time we need that package. To check the library do following commands:
(new-env) C:\Users\admin>python
Python 3.6.8 (tags/v3.6.8:3c6b436a57, Dec 24 2018, 00:16:47) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> cv2.__version__
'3.4.2'

Conclusions

Two ways to install Python and OpenCV library. The first way quite hard and inefficient. The second on is recommendation for separated project.

References

[1]. https://sebastianraschka.com/Articles/2014_python_2_3_key_diff.html

[2]. https://en.wikipedia.org/wiki/Anaconda_(Python_distribution)

[3]. https://en.wikipedia.org/wiki/R_(programming_language)

[4]. https://anaconda.org/conda-forge/opencv