HARK-OPENCV provides OpenCV-based image processing in HARK.


We provide multiple object tracking and feature extraction which are perfectly compatible with HARK audio processing. The benefits are as follows.

  1. Provides the latest image processing based on OpenCV to HARK
  2. GUI based image processing programing
  3. Integration of image-based multiple object tracking and audio-based sound source tracking
  4. Integration of an image-based feature and an audio-based feature
  5. Perfect compatibility with HARK audio processing

HARK-OPENCV Functionalities


All the following functionalities are available in HARK

  • Image I/O
  • Image filtering such as binarization, etc.
  • Image-based object segmentation such as Hough circle detection, etc.
  • Image-based multiple object tracking for each segments
  • Image-based feature extraction for each segments such as surf, etc.
  • Data transformation for integrating with HARK audio processing results
  • Image communication with ROS

HARK-OPENCV Installation Instructions

Go to following page.



Go to following page.



HARK-OPENCV Installation Instructions



Any version later than OpenCV-1.0 is OK for HARK-OPENCV.
See OpenCV official site for the detail.

  • Example in Ubuntu 10.04
sudo aptitude install libcv4 libhighgui4 libcvaux4 libcv-dev libhighgui-dev libcvaux-dev python-opencv opencv-doc


  • harkfd (for HARK main software)
  • hark-opencv (optional modules for the OpenCV)

See next subsection for the installation detail.


Installation from pre-compiled binaries

Compilation Environment

  • OS
    • See HARK Installation Instructions.

For other environments, please go for the source compilation.

The pre-compiled binaries are supported up to HARK version 2.2.0.
To use HARK later than version 2.2.0, go for the source compilation.



For adding the HARK repository, see HARK Installation Instructions.

HARK stack (harkfd, hark-opencv) installation

sudo apt-get install harkfd hark-opencv


Installation from source compilation (Optional)

Supported Environment

  • OS
    • 32bit/64bit Ubuntu 10.04 or later


Step1 : OpenCV Installation

Fistly, install OpenCV fundamental packages before you proceed.

      • See OpenCV official site for the detail.


Step2 : HARK Basic Package Installation

For adding the HARK repository, see HARK Installation Instructions.

Hereinafter, the source compilation of only HARK-OPENCV package is explained.
We assume that the HARK basic packages (such as harkfd, flowdesigner-0.9.1-hark) are installed by pre-compiled bineries.

sudo apt-get install harkfd


If you use the Ubuntu 64bit version, you have to install the HARK basic packages by source compilation as well.

HARK-OPENCV package is installed by source compilation in the next subsection.


Step3 : HARK-OPENCV Package Installation

To install HARK-OPENCV, we need source codes of both hark-opencv and harkfd.

Firstly, move to your preferred directory to download source codes.



Download the latest hark-opencv-*.*.*.tar.gz from here and unzip them.

tar zxvf hark-opencv-*.*.*.tar.gz


Compile and install the source codes.

cd hark-opencv-*.*.*
./configure --prefix=/usr/local --with-hark-inc=/usr/include/hark; make clean; make ; sudo make install;


The “*.*.*” describes the package version.

Here, we need header files of harkfd with –with-hark-inc option.

If you haven’t set FLOWDESIGNER_PATH in ~/.bashrc, use the following command to activate this package.

echo "export FLOWDESIGNER_PATH=$FLOWDESIGNER_PATH:/usr/local/lib" >> ~/.bashrc


Check the installation

To check the HARK-OPENCV installation, follow this section.

Run HARK-designer in a new terminal.



Then, you will find a browser is opend.
HARK-OPENCV modules can be found at the following place.

Preferences > Packages > hark-opencv-cv.def, hark-opencv.def


Preferences > Version > HARK_OPENCV *.*.* (Revision:*)

Back to Top