I had previously used OpenCV 2.4.5 with some certain configs and packages on Ubuntu 12.04.1 but had issues upgrading to OpenCV 2.4.6.1 on Ubuntu 12.04.2

I would like to share some ideas (a compilation of noteworthy information gathered from several sources including SO, ubuntu.org, asklinux.org and many other; and of course by trying several procedures)

Below is what eventually got me through.

NOTE: ensure you uninstall any previous installation of OpenCV, FFMpeg and other dependencies previously installed.

STEP 1 (install ffmpeg and dependencies)

# goto http://sourceforge.net/projects/opencvlibrary/files/opencv-unix/
# download the latest stable opencv such as 2.4.6.1 (http://sourceforge.net/projects/opencvlibrary/files/opencv-unix/2.4.5/opencv-2.4.5.1.tar.gz/download) to current directory (such as home or ~/Document)
# cd /opt
# tar -xvf <path-to-download-folder>/OpenCV-2.4.6.1.tar.gz
# cd OpenCV-2.4.6.1
# create a foler under current dir (following previous step, this should be <opencv-dir>), called prepare
# cd prepare
# Copy the following script to gedit and save as install.sh to current dir, this should be <opencv-dir>/prepare
# Check corresponding url used in the script for latest versions of the package and replace as required
# Open terminal and navigate to location used above
# sudo chmod +x install.sh
# ./install

echo "Removing any pre-installed ffmpeg, x264, and other dependencies (not all the previously installed dependecies)"
sudo apt-get remove ffmpeg x264 libx264-dev libvpx-dev librtmp0 librtmp-dev libopencv-dev
sudo apt-get update

arch=$(uname -m)
if [ "$arch" == "i686" -o "$arch" == "i386" -o "$arch" == "i486" -o "$arch" == "i586" ]; then
flag=0
else
flag=1
fi

echo "Installing Dependenices"
sudo apt-get install autoconf automake make g++ curl cmake bzip2 python unzip \
  build-essential checkinstall git git-core libass-dev libgpac-dev \
  libsdl1.2-dev libtheora-dev libtool libva-dev libvdpau-dev libvorbis-dev libx11-dev \
  libxext-dev libxfixes-dev pkg-config texi2html zlib1g-dev

echo "downloading yasm (assembler used by x264 and FFmpeg)"
# use git or tarball (not both)
wget http://www.tortall.net/projects/yasm/releases/yasm-1.2.0.tar.gz
tar xzvf yasm-1.2.0.tar.gz
cd yasm-1.2.0

echo "installing yasm"
./configure
make
sudo make install
cd ..

echo 'READ NOTE BELOW which was extracted from http://wiki.serviio.org/doku.php?id=build_ffmpeg_linux'
echo 'New version of x264 contains by default support of OpenCL. If not installed or without sense (example Ubuntu 12.04LTS on VMWare) add to configure additional option --disable-opencl. Without this option ffmpeg could not be configured (ERROR: libx264 not found).'

echo "downloading x264 (H.264 video encoder)"
# use git or tarball (not both)
# git clone http://repo.or.cz/r/x264.git or
git clone git://git.videolan.org/x264.git
cd x264
# wget ftp://ftp.videolan.org/pub/videolan/x264/snapshots/x264-snapshot-20130801-2245-stable.tar.bz2
# tar -xvjf x264-snapshot-20130801-2245-stable.tar.bz2
# cd x264-snapshot-20130801-2245-stable/

echo "Installing x264"
if [ $flag -eq 0 ]; then
./configure --enable-static --disable-opencl
else
./configure --enable-shared --enable-pic --disable-opencl
fi
make
sudo make install
cd ..

echo "downloading fdk-aac (AAC audio encoder)"
# use git or tarball (not both)
git clone --depth 1 git://github.com/mstorsjo/fdk-aac.git
cd fdk-aac

echo "installing fdk-aac"
autoreconf -fiv
./configure --disable-shared
make
sudo make install
cd ..

echo "installing libmp3lame-dev (MP3 audio encoder.)"
sudo apt-get install libmp3lame-dev

echo "downloading libopus (Opus audio decoder and encoder.)"
wget http://downloads.xiph.org/releases/opus/opus-1.0.3.tar.gz
tar xzvf opus-1.0.3.tar.gz
cd opus-1.0.3

echo "installing libopus"
./configure --disable-shared
make
sudo make install
cd ..

echo "downloading libvpx VP8/VP9 video encoder and decoder)"
# use git or tarball (not both)
git clone --depth 1 http://git.chromium.org/webm/libvpx.git
cd libvpx
# wget http://webm.googlecode.com/files/libvpx-v1.1.0.tar.bz2 (this seems not to be update, but can still be used if the fedoraproject link below is not available))
# wget http://pkgs.fedoraproject.org/repo/pkgs/libvpx/libvpx-v1.2.0.tar.bz2/400d7c940c5f9d394893d42ae5f463e6/libvpx-v1.2.0.tar.bz2
# tar xvjf libvpx-v1.2.0.tar.bz2
# cd libvpx-v1.2.0

echo "installing libvpx"
./configure --disable-examples
make
sudo make install
cd ..

sudo ldconfig

echo "downloading ffmpeg"
# git clone http://repo.or.cz/r/ffmpeg.git
git clone git://source.ffmpeg.org/ffmpeg.git
cd ffmpeg/
# wget http://ffmpeg.org/releases/ffmpeg-2.0.tar.bz2
# tar -xvjf ffmpeg-2.0.tar.bz2
# cd ffmpeg-2.0/

echo "installing ffmpeg" 
if [ $flag -eq 0 ]; then
./configure --enable-gpl --enable-libass --enable-libfdk-aac --enable-libopus --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvorbis --enable-libx264 --enable-libxvid --enable-nonfree --enable-postproc --enable-version3 --enable-x11grab --enable-libvpx
else
./configure --enable-gpl --enable-libass --enable-libfdk-aac --enable-libopus --enable-libfaac --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libtheora --enable-libvorbis --enable-libx264 --enable-libxvid --enable-nonfree --enable-postproc --enable-version3 --enable-x11grab --enable-libvpx --enable-shared
fi

make
sudo make install
hash -r

cd .. # move up one level to prepare folder
cd .. # move up one level to opencv folder

echo "Checking to see if you're using your new ffmpeg"
ffmpeg 2>&1 | head -n1

sudo ldconfig

STEP 2 (Install OpenCV and necessary packages)

echo "Installing Dependenices"    
sudo apt-get install libtiff4-dev libjpeg-dev libjasper-dev

echo "installing Video I/O libraries, support for Firewire video cameras and video streaming libraries"
sudo apt-get install libav-tools libavcodec-dev libavformat-dev libswscale-dev libdc1394-22-dev libxine-dev libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev libv4l-dev v4l-utils v4l-conf

echo "installing the Python development environment and the Python Numerical library"
sudo apt-get install python-dev python-numpy

echo "installing the parallel code processing library (the Intel tbb library)"
sudo apt-get install libtbb-dev

echo "installing the Qt dev library"
sudo apt-get install libqt4-dev libgtk2.0-dev

echo "installing other dependencies (if need be it would upgrade current version of the packages)"
sudo apt-get install patch subversion ruby librtmp0 librtmp-dev libfaac-dev libmp3lame-dev libopencore-amrnb-dev libopencore-amrwb-dev libvpx-dev libxvidcore-dev

echo installing optional packages"
sudo apt-get install libdc1394-utils libdc1394-22-dev libdc1394-22 libjpeg-dev libpng-dev libtiff-dev libjasper-dev

STEP 3 (run ldconfig)

# Open a new terminal window
# Open /etc/ld.so.conf and check, 
# if the paths "/usr/lib" and "/usr/local/lib" including the quote exist in the file. If not, add them manually or by
    sudo echo "/usr/local/lib" >> /etc/ld.so.conf
    sudo echo "/usr/lib" >> /etc/ld.so.conf
# execute the following
    sudo ldconfig

STEP 4a (Build & Install for OS Usage)

# still ensure you haven't close the new terminal window open in STEP 3
# execute the following
mkdir os_build
cd os_build
cmake -DCMAKE_BUILD_TYPE=RELEASE -DCMAKE_INSTALL_PREFIX=/usr/local -DBUILD_NEW_PYTHON_SUPPORT=ON -DINSTALL_PYTHON_EXAMPLES=ON -DWITH_TBB=ON -DWITH_V4L=ON -DINSTALL_C_EXAMPLES=ON -DBUILD_EXAMPLES=ON -DWITH_QT=ON -DWITH_OPENGL=ON -DWITH_OPENCL=ON -DWITH_EIGEN=ON -DWITH_OPENEXR=ON ..

    make
    sudo make install

# add the following to user environment variable ~/.bashrc
    export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/usr/local/lib
    export PKG_CONFIG_PATH=${PKG_CONFIG_PATH}:/usr/local/lib/pkgconfig
# execute the following
    sudo ldconfig
# start to use and enjoy opencv, it should have been install into any of these locations
#   /usr/local/include/opencv2, /usr/local/include/opencv, /usr/include/opencv, /usr/include/opencv2, /usr/local/share/opencv
#   /usr/local/share/OpenCV, /usr/share/opencv, /usr/share/OpenCV, /usr/local/bin/opencv*, /usr/local/lib/libopencv*

STEP 4b (Build for Java Usage): OPTIONAL

# still ensure you haven't close the new terminal window open in STEP 4
# execute the following
    cd ..
    mkdir java_build
    cd java_build
    cmake -DCMAKE_BUILD_TYPE=RELEASE -DBUILD_SHARED_LIBS=OFF -DINSTALL_PYTHON_EXAMPLES=ON -DWITH_TBB=ON -DWITH_V4L=ON -DINSTALL_C_EXAMPLES=ON -DBUILD_EXAMPLES=ON -DWITH_QT=ON -DWITH_OPENGL=ON -DWITH_OPENCL=ON -DWITH_EIGEN=ON -DWITH_OPENEXR=ON ..

    make

# You can check the "java_build/bin" directory to locate the jar and libopencv_java.so file for your development
# As stated in the docs, the Java bindings dynamic library is all-sufficient, i.e. doesn’t depend on other OpenCV libs, but includes all the OpenCV code inside

STEP 5 (install v4l: Note: installing v4l-utils after opencv installation works for Ubuntu 12.04.2 & OpenCV 2.4.6.1)

# still ensure you haven't close the new terminal window open in STEP 3
# goto http://www.linuxtv.org/downloads/v4l-utils
# download the latest v4l such as v4l-utils-0.9.5.tar.bz2
# copy the downloaded file to the current terminal dir (following previous step, this should be <opencv-dir>/prepare)
# execute the following
    tar -xvjf v4l-utils-0.9.5.tar.bz2
    cd v4l-utils-0.9.5/
    ./configure
    make
    sudo make install
    cd ..
    cd .. # (to go to <opencv-dir>)
    sudo ldconfig

Worth Noting

# To check the path where opencv & other lib files are stored, do:


pkg-config --cflags opencv

    (output will come as)
    -I/usr/include/opencv



pkg-config --libs opencv

    (output will come as)
    -lopencv_core -lopencv_imgproc -lopencv_highgui -lopencv_ --ml -lopencv_video -lopencv_features2d -lopencv_calib3d -lopencv_objdetect -lopencv_contrib -lopencv_legacy -lopencv_flann

# The above paths are needed to compile your opencv programs, as given in the next illustration.

# write a simple C program to test, by saving below program in a file named DisplayImage.c

#include <stdio.h>
    #include <opencv2/highgui/highgui.hpp>

    int main(int argc, char *argv[]) {
        IplImage* img=0; /* pointer to an image */
        printf("Hello\n");

        if(argv[1] != 0)
            img = cvLoadImage(argv[1], 0); // 1 for color
        else
            printf("Enter filename\n");

        if(img != 0) {
            cvNamedWindow("Display", CV_WINDOW_AUTOSIZE); // create a window
            cvShowImage("Display", img); // show image in window
            cvWaitKey(0); // wait until user hits a key
            cvDestroyWindow("Display");
        }
        else
            printf("File not found\n");

        return 0;
    }

# write a simple C++ program to test, by saving below program in a file named DisplayImage.cpp

#include <stdio.h>
#include <opencv2/opencv.hpp>
#include <opencv2/highgui/highgui.hpp>

using namespace cv;

int main( int argc, char** argv )
{
  Mat image;
  image = imread( argv[1], 1 );

  if( argc != 2 || !image.data )
    {
      printf( "No image data \n" );
      return -1;
    }

  namedWindow( "Display Image", CV_WINDOW_AUTOSIZE );
  imshow( "Display Image", image );

  waitKey(0);

  return 0;
}

# To compile & run :



g++ <filename: e.g. DisplayImage.c or DisplayImage.cpp> `pkg-config --cflags --libs opencv` && ./a.out img

or



g++ -I/usr/include/opencv -I/usr/local/include -lopencv_core -lopencv_highgui -lopencv_ml -lopencv_imgproc -lopencv_video -lopencv_features2d -lopencv_calib3d -lopencv_objdetect -lopencv_contrib -lopencv_legacy -lopencv_flann -lopencv_nonfree <filename: e.g. DisplayImage.c or DisplayImage.cpp> && ./a.out img

where “img” is the name of any image with extension within the same folder .
You should be able to see “Hello” and the image in a different window.

If this runs, Congrats! now you can run any C/C++ program with opencv lib.


# Now lets simplify the above big command by making a shortcut for it:
go to your local home directory(cd /home/) and open the .bashrc file using gedit(the file will be hidden). Append the following to the file:



alias gcv="g++ -I/usr/include/opencv -I/usr/local/include -lopencv_core -lopencv_highgui -lopencv_ml -lopencv_imgproc -lopencv_video -lopencv_features2d -lopencv_calib3d -lopencv_objdetect -lopencv_contrib -lopencv_legacy -lopencv_flann -lopencv_nonfree"

and save. Close the terminal and open it again.(as this process requires relogin of the terminal)

# Now, go to directory containing a sample program & do



gcv <filename: e.g. DisplayImage.c or DisplayImage.cpp> && ./a.out

or



gcv <filename: e.g. DisplayImage.c or DisplayImage.cpp>
    ./a.out input_img.jpg

As you can see the commands now become similar to $cc filename.c, $./a.out which are used normally for compiling and executing C/C++ programs.

Some ways to check whether all lib files are installed-

apt-cache search opencv

returns:

libcv-dev - Translation package for libcv-dev
libcv2.3 - computer vision library - libcv* translation package
libcvaux-dev - Translation package for libcvaux-dev
libcvaux2.3 - computer vision library - libcvaux translation package
libhighgui-dev - Translation package for libhighgui-dev
libhighgui2.3 - computer vision library - libhighgui translation package
libopencv-calib3d-dev - development files for libopencv-calib3d
libopencv-calib3d2.3 - computer vision Camera Calibration library
libopencv-contrib-dev - development files for libopencv-contrib
libopencv-contrib2.3 - computer vision contrib library
libopencv-core-dev - development files for libopencv-core
libopencv-core2.3 - computer vision core library
libopencv-dev - development files for opencv
libopencv-features2d-dev - development files for libopencv-features2d
libopencv-features2d2.3 - computer vision Feature Detection and Descriptor Extraction library
libopencv-flann-dev - development files for libopencv-flann
libopencv-flann2.3 - computer vision Clustering and Search in Multi-Dimensional spaces library
libopencv-gpu-dev - development files for libopencv-gpu
libopencv-gpu2.3 - computer vision GPU Processing library
libopencv-highgui-dev - development files for libopencv-highgui
libopencv-highgui2.3 - computer vision High-level GUI and Media I/O library
libopencv-imgproc-dev - development files for libopencv-imgproc
libopencv-imgproc2.3 - computer vision Image Processing library
libopencv-legacy-dev - development files for libopencv-legacy
libopencv-legacy2.3 - computer vision legacy library
libopencv-ml-dev - development files for libopencv-ml
libopencv-ml2.3 - computer vision Machine Learning library
libopencv-objdetect-dev - development files for libopencv-objdetect
libopencv-objdetect2.3 - computer vision Object Detection library
libopencv-video-dev - development files for libopencv-video
libopencv-video2.3 - computer vision Video analysis library
opencv-doc - OpenCV documentation and examples
python-opencv - Python bindings for the computer vision library

回答1:

Thanks for the information. I tried to run these commands but encountered several problems. I am using a brand new Ubuntu 12.04 install.

1) The first apt-get remove command asked me if I wanted to remove a lot of things that I wasn't comfortable removing, like LibreOffice. I didn't remove anything in this step, which I figured would be okay, since the only program I've installed is Java SDK and JRE. Why did you opt for such a large list of programs to remove?

2) Adding sudo to the git and tar commands was necessary.

3) When installing opus and ffmpeg, there were many error messages for the ./configure command because of a lack of permission for writing and ultimately the configurations failed.

Anyways, just thought I would log these problems on the post in the hopes that it will improve the procedure listed above.

回答2:

To install OpenCV on Ubuntu from http://namhuy.net/1205/how-to-install-opencv-on-ubuntu.html

$ su -
# apt-get update
# apt-get install build-essential
# apt-get install libavformat-dev
# apt-get install x264 v4l-utils ffmpeg
# apt-get install libcv2.3 libcvaux2.3 libhighgui2.3 python-opencv opencv-doc libcv-dev libcvaux-dev libhighgui-dev

回答3:

Here is my blog post on installing OpenCV with it's many dependencies ie Cuda, OpenNI, Gstreamer, ffmpeg, QT5, ... ect. http://drewski1992.blogspot.com/2014/04/install-opencv-24x-with-cuda-55-or-6.html

Following this should give you a full OpenCV build.

回答4:

thanks for the info. I installed everything in Ubuntu 13.04 and ran into a problem when trying to use the gcv shortcut. Using the

g++ `pkg-config --cflags --libs opencv`

shortcut works perfectly. Previously I had to link all the libraries manually in a Makefile so this is still much better. Appreciate you sharing the info, thanks!