Compile Latest OpenCV on Ubuntu

Compiling OpenCV can vary depending on the platform. In this demonstration, I will be using Ubuntu. However, these steps are mostly the same for any other distribution, with the only differences being the package manager and the package names.

Retrieve OpenCV source code

This shell command retrieve the most recent OpenCV version number from the GitHub API and store it in a shell variable for future use.

vernum=$(wget -qO- -t1 -T2 "https://api.github.com/repos/opencv/opencv/releases/latest"  grep "tag_name"  head -n 1  awk -F ":" '{print $2}'  sed 's/\"//g;s/,//g;s/ //g;s/v//g')

Next, download and unzip the OpenCV source code from GitHub. To avoid cluttering the working directory, it’s recommended to create a new folder specifically for the compilation process.

mkdir opencv_build && cd opencv_build

wget https://github.com/opencv/opencv/archive/refs/tags/$vernum.zip
unzip $vernum.zip && rm $vernum.zip

wget https://github.com/opencv/opencv_contrib/archive/refs/tags/$vernum.zip
unzip $vernum.zip && rm $vernum.zip

cd opencv-$vernum

Install dependencies

Before proceeding, make sure to update your system.

sudo apt update && sudo apt upgrade

The commands for installing the following packages include libraries that are commonly used with OpenCV, such as v4l, FFMPEG, and x264, but do not include GPU acceleration or machine learning components.

sudo add-apt-repository "deb http://security.ubuntu.com/ubuntu xenial-security main"
sudo apt install -y build-essential cmake git unzip zip\
                pkg-config libavcodec-dev libavformat-dev libswscale-dev \
                libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev \
                libsnappy-dev libboost-all-dev libjasper-dev libdc1394-22-dev\
                python3-dev python3-numpy python3-pip ffmpeg libopenblas-dev \
                tesseract-ocr libtesseract-dev libprotobuf-dev libleveldb-dev \
                libhdf5-serial-dev protobuf-compiler libatlas-base-dev \
                libgflags-dev libgoogle-glog-dev liblmdb-dev libfaac-dev \
                gfortran libgstreamer1.0-dev libatlas-base-dev libxvidcore-dev \
                libpng-dev libopenexr-dev libwebp-dev \
                libmp3lame-dev libtheora-dev libvorbis-dev  \
                libopencore-amrwb-dev x264 v4l-utils libgdk-pixbuf2.0-dev \
                manpages-dev libopencore-amrnb-dev \
                libgstreamer-plugins-base1.0-dev libavresample-dev 
pip3 install --upgrade pip
apt-get install -f

Configure CMake project

Remenber to execute all the build commands within a temporary directory. This way, if you make a mistake, you can easily delete the entire temporary folder to quickly start over.

mkdir build && cd build

To display a list of all CMake compile options, run the following command:

cmake .. -LAH

L List cached variables.
A Show options marked as advanced.
H Show more helpful information about each option.

Here’s an example configure command that you can use. If you’ve successfully installed the dependencies according to the previous instructions, this example cmake command should work for you.

cmake -D CMAKE_BUILD_TYPE=RELEASE \
        -D CMAKE_INSTALL_PREFIX=/usr/local/ \
        -D OPENCV_ENABLE_NONFREE=ON \
        -D OPENCV_EXTRA_MODULES_PATH=~/opencv_build/opencv_contrib-$vernum/modules \
        -D OPENCV_GENERATE_PKGCONFIG=YES \
        -D WITH_QT=OFF \
        -D WITH_OPENGL=OFF \
        -D WITH_CUDA=OFF \
        -D BUILD_EXAMPLES=ON \
        -D INSTALL_PYTHON_EXAMPLES=ON \
        -D INSTALL_C_EXAMPLES=ON ..

If you find it overwhelming to deal with extensive text in the command line and have a GUI installed on your machine, consider trying tools like ccmake and cmake-gui for a more user-friendly experience.

You can also check an existed OpenCV build what compile flags are used by a function built into OpenCV. In C++ it looks like this:
cv::getBuildInformation()
In Python:
cv2.getBuildInformation()

Compile

We’ve completed the necessary steps, and now it’s in the hands of your CPUs. The -j flag specifies how many threads will be used for compilation, and it should not exceed the number of CPU cores you have.

make -j6

If the compilation process proceeds without errors, you can proceed with the installation by simply running this single line of code:

make install

At this point, CMake should be able to detect the OpenCV library files.

deemoe's blog