Compile OpenCV3 on Raspberry Pi

Why

The default version installed using "sudo apt-get" is 2.4. Therefore, we need compile OpenCV3+ using source code.

Platform

raspberry pi (linux kernel)

Prepare

pi@raspberrypi:~/opencv-3.3.1 $ sudo apt-get install cmake-gui wget

How

Download the OpenCV source code

cd ~
wget -O opencv.zip https://github.com/Itseez/opencv/archive/3.3.1.zip
unzip opencv.zip
wget -O opencv_contrib.zip https://github.com/Itseez/opencv_contrib/archive/3.3.1.zip
unzip opencv_contrib.zip

Install OpenCV on System

pi@raspberrypi:~/opencv-3.3.1 $ cmake-gui  &

Click "Generate" after "Configure" completed.

make:

pi@raspberrypi:~/opencv-3.3.1/build $ sudo make & sudo make install

Install OpenCV using virtual environment on raspberry Pi

Create your Python virtual environment and install NumPy

We’ll be using Python virtual environments, a best practice when working with Python.

You can install pip, virtualenv, and virtualenvwrapper using the following commands:

wget https://bootstrap.pypa.io/get-pip.py
sudo python get-pip.py
sudo python3 get-pip.py
sudo pip install virtualenv 
sudo pip install virtualenvwrapper
sudo rm -rf ~/.cache/pip

Once both virtualenv  and virtualenvwrapper  have been installed, open up your~/.profile  and append the following lines to the bottom of the file, using your favorite terminal-based text editor such as vim , emacs , or nano :

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

From there, reload your ~/.profile  file to apply the changes to your current bash session:

source ~ / .profile

You’ll need to run source ~/.profile  each time you open a new terminal/SSH into your Pi to ensure your system variables have been set properly (it also loads this file on boot).

Next, create your Python 3 virtual environment:

mkvirtualenv cv -p python3

Here I am creating a Python virtual environment named cv  using Python 3 (alternatively, you may also use Python 2.7 by changing the -p  switch to  python2 ).

You can name the virtual environment whatever you want, but I use cv  as the standard naming convention here on PyImageSearch.

Finally, install NumPy into the Python virtual environment:

pip install numpy

Compile and install the optimized OpenCV library for Raspberry Pi

We’re now ready to compile and install the optimized version of Raspberry Pi.

Ensure you are in the cv  virtual environment using the workon  command:

workon cv

You can use 'deactivate' to exit python virtual environment:

deactivate

Here I use opencv-3.3.0. 

$ cd ~/opencv-3.3.0/
$ mkdir build
$ cd build
$ cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.3.0/modules \
    -D ENABLE_NEON=ON \
    -D ENABLE_VFPV3=ON \
    -D BUILD_TESTS=OFF \
    -D INSTALL_PYTHON_EXAMPLES=OFF \
    -D BUILD_EXAMPLES=OFF ..

Notice how the NEON and VFPV3 flags have been enabled.

If you’re compiling OpenCV for Python 3, check the “Python 3” output of CMake: (If you use Python Virtual Environment 'cv')

Notice how the Interpreter , Libraries , numpy , and packages path  variables have been properly set.

Before you start the compile I would suggest increasing your swap space. This will enable you to compile OpenCV with all four cores of the Raspberry Pi without the compile hanging due to memory exhausting.

Open up your /etc/dphys-swapfile  file and then edit the CONF_SWAPSIZE  variable:

CONF_SWAPSIZE=1024

Notice that I’m increasing the swap from 100MB to 1024MB. This is the secret sauce to compiling OpenCV with multiple cores on the Raspbian Stretch.

If you do not perform this step it’s very likely that your Pi will hang.

From there, restart the swap service:

sudo /etc/init.d/dphys-swapfile restart

Note: Increasing swap size is a great way to burn out your Raspberry Pi microSD card. Flash-based storage have limited number of writes you can perform until the card is essentially unable to hold the 1’s and 0’s anymore. We’ll only be enabling large swap for a short period of time, so it’s not a big deal. Regardless, be sure to backup your  .img  file after installing OpenCV + Python just in case your card dies unexpectedly early. You can read more about large swap sizes corrupting memory cards on this page.

Assuming OpenCV compiled without error (as in my screenshot above), you can install your optimized version of OpenCV on your Raspberry Pi:

make -j2

Note: 

    System reboot when I use "make -j4" on raspberry 3. I also tried "make" on raspberry-zero board. It took about half a day to compile. Therefore, it's very useful to use multiple job tasks and this will save a lot of time.

Error:

[ 65%] Linking CXX static library ../../lib/libopencv_ts.a
[ 65%] Built target opencv_ts
Makefile:160: recipe for target 'all' failed
make: *** [all] Error 2

I solved this error in this way:

• make clean
• use Cmake to generate Makefile
• make (not make -jx)

Of course, this will take a longer time to make.

Assuming OpenCV compiled without error (as in my screenshot above), you can install your optimized version of OpenCV on your Raspberry Pi:

 sudo make install

sudo ldconfig

Don’t forget to go back to your /etc/dphys-swapfile  file and:

1. Reset CONF_SWAPSIZE  to 100MB.
2. Restart the swap service.

Testing your optimized OpenCV + Raspberry Pi install

I installed OpenCV directly on OS, rather than Python Virtual Environment.

Python2.7:

pi@raspberrypi:~/opencv-3.3.0/build $ python
Python 2.7.13 (default, Nov 24 2017, 17:33:09)
[GCC 6.3.0 20170516] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> quit()

Python3:

pi@raspberrypi:~/opencv-3.3.0/build $ python3
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
[GCC 6.3.0 20170124] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named 'cv2'
>>> quit()

If you compiled OpenCV for Python 3, you need to issue the following commands to sym-link the cv2.so  bindings into Python3:

pi@raspberrypi:~/opencv-3.3.0/build $ cd /usr/local/lib/py
pypy2.7/   python2.7/ python3.5/
pi@raspberrypi:~/opencv-3.3.0/build $ cd /usr/local/lib/python3.5/site-packages/
pi@raspberrypi:/usr/local/lib/python3.5/site-packages $ ls
cv2.cpython-35m-arm-linux-gnueabihf.so
pi@raspberrypi:/usr/local/lib/python3.5/site-packages $ sudo mv cv2.cpython-35m-arm-linux-gnueabihf.so cv2.so

pi@raspberrypi:/usr/local/lib/python3.5/site-packages $ python3
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
[GCC 6.3.0 20170124] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> quit()

Test OpenCV with 'cv' Python3 virtual environment:

(cv) pi@raspberrypi:~ $ python
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
[GCC 6.3.0 20170124] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named 'cv2'
>>>

If you compiled OpenCV for Python 3, you need to issue the following commands to sym-link the cv2.so  bindings into your cv  virtual environment:

(cv) pi@raspberrypi:~ $ cd ~/.virtualenvs/cv/lib/python3.5/site-packages/

(cv) pi@raspberrypi:~/.virtualenvs/cv/lib/python3.5/site-packages $ ln -s /usr/local/lib/python3.5/site-packages/cv2.so cv2.so
(cv) pi@raspberrypi:~/.virtualenvs/cv/lib/python3.5/site-packages $ cd ~
(cv) pi@raspberrypi:~ $ python
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
[GCC 6.3.0 20170124] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cv2
>>> cv2.__version__
'3.3.0'

Congratulations!!!

Reference:

1. optimizing-opencv-on-the-raspberry-pi, https://www.pyimagesearch.com/2017/10/09/optimizing-opencv-on-the-raspberry-pi/
2.