Installing OpenCV

By dreamcaster on 15:50 Apr.26-2017

Installing OpenCV on Windows 10 can be a bit tricky. Nothing too fancy, but yes, it can be a bit tiresome, especially for the 3rd party libraries.

Here is my walkthrough.

CMake

Install CMake via the MSI

$ cmake --version
cmake version 3.8.0

CMake suite maintained and supported by Kitware (kitware.com/cmake).

Clone the Repo

Obtain the OpenCV repo from GitHub:

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

set OPENCV_REPO=D:\pub\opencv

Install Python, Pip and Numpy

All this weird python jargon is a bit tiresome, but amounts nothing much than to install a few required modules with the usual command-line dependency manager.

$ python --version
Python 3.6.1

$ pip --version
pip 9.0.1 from d:\python\python36\lib\site-packages (python 3.6)

$ pip install numpy
Collecting numpy
  Downloading numpy-1.12.1-cp36-none-win_amd64.whl (7.7MB)
Installing collected packages: numpy
Successfully installed numpy-1.12.1

$ pip freeze
numpy==1.12.1

Intel Libraries

Download the Windows versions for these babies to %OPENCV_REPO%\deps

Intel(R) Threading Building Blocks 2017 Update 5

[Intel® Integrated Performance Primitives] (https://software.intel.com/en-us/intel-ipp) beware, this library is huge, and requires registration

Intel IPP Asynchronous C/C++ mentioned at the base tutorial is unavailable at the moment. Apparently, it has been withdrawn from public preview by Intel more than two years ago.