Introduction

What is OpenCV?

OpenCV (Open Source Computer Vision) is a library of programming functions mainly aimed at real time computer vision.

Example applications of the OpenCV library are Human-Computer Interaction (HCI); Object Identification, Segmentation and Recognition; Face Recognition; Gesture Recognition; Motion Tracking, Ego Motion, Motion Understanding; Structure From Motion (SFM); and Mobile Robotics.

How do I Download OpenCV?

You can either download the last published version for Linux or Windows. Or you can try the CVS version of OpenCV

Installing OpenCV

1. Installing OpenCV on Linux

Getting Started with OpenCV

1. Using Visual C++ and OpenCV

2. Using Eclipse IDE and OpenCV

3. Using C++ Builder IDE and OpenCV

4. Using DevCpp IDE and OpenCV

5. C/C++ interpreter Ch and OpenCV.

6. How do I compile OpenCV and OpenCV programs in Linux? - Also see notes in the Linux section of OS Specific Stuff below

7. Using OpenCV with Visual C++ and Microsoft's DirectShow

Frequently Asked Questions

Problem Specific Frequently Asked Questions

1. Beginners/Newbie FAQ's (how to access a pixel etc.)

2. How to use OpenCV to capture and display images from a camera

3. Which video file formats can I use with OpenCV?

4. How do I use OpenCV to Detect Faces?

5. Using OpenCV for Face Recognition

6. How to use OpenCV to Detect Objects?

7. How to use the OpenCV Blob Extraction Library

8. How do I correctly refresh my HighGUI application?

9. Where to download DirectShow?

10. How to Display more than one image in a single window?

11. CodingStyleGuide

12. VideoSurveillance

Tutorials

1. Using OpenCV with Microsoft's DirectShow

2. POSIT tutorial (wiki page incomplete but in progress)

3. A short introduction to programming with OpenCV

4. Tutorial on how to use Philips webcam features with OpenCV on Linux

5. Introduction to OpenCV (in Turkish)

6. Introduction to programming with OpenCV and IPP - (2003)

7. Sample program for Face Recognition

Support, Documentation & Resources

Technical Support

There is no formal technical support for OpenCV, but there is a very active OpenCV Forum at Yahoo Groups where you are welcome to post your questions.

Documentation of Functions

• CxCore

• CvReference

• CvAux

• HighGui

OpenCV Online Resources

• The Official Home Page

• The OpenCV China (OpenCV中文站)

• The SourceForge main page: http://sourceforge.net/projects/opencvlibrary - Note, the bug and patch trackers on sourceforge are not regularly checked. Submit bugs and patches to the OpenCV Yahoo Group.

• The OpenCV Forum at Yahoo Groups

• The #opencv IRC channel on Freenode -- currently 'Un-official'

• OpenCVDotNet - OpenCV wrappers for .NET Framework.

• SoftIntegration - OpenCV wrappers for Ch.

• OpenCV Naver Community Forum (Korean)

• EffortFor3dReconstruction - Community document to bring forward 3D reconstruction in OpenCV

Ideas for the Future of OpenCV

• Our SummerOfCode page describes project ideas for the Google Summer of Code. The SummerOfCodeApplication page shows current status of our application to the program.

• Mac_OS_X_OpenCV_Port holds specific ideas for OpenCV on Mac OS X

OS Specific Stuff

Linux Version

• In beta 5, test programs are not built by default. To build, run "make check" (see CxTest for more information on running and writing tests).

• The current distribution of OpenCV (0.9.7 beta 5) does *NOT* work with GCC 4.0 or 4.1. Before compiling, you will have to run the configure script as follows: CXXFLAGS=-fno-strict-aliasing ./configure . GCC 4.0 will *NOT* (but 4.1 will) give *ANY* warning whatsoever that something is wrong (it is a bug of GCC 4.0), but it will *NOT* work, so you have been warned. If this option is not used with GCC 4.0 or 4.1, the resulting library has lots of errors as evidenced by failed tests in the test programs and the test programs segfaulting. Alternatively, you can also use the CVS version. The aliasing bugs are fixed in the CVS version.

• Building will fail on x86_64 platforms and GCC 4.0.2 using the beta 5 release with errors about various SSE things like "'__m128d' was not declared" . Since GCC 4 and beta 5 are broken anyway, use the CVS version. GCC 3.2.3 may also work.

• Sometimes the configure script may generate a bad makefile for your system if it was generated by a version of autoconf/automake/etc that is too old for your system. For example, you may get linking errors for the library files on x86_64. To fix this, run autoreconf in the root opencv directory (where configure is) to regenerate the scripts. Then run make distclean and run the newly generated configure to generate new makefiles.

• Compiling OpenCV and OpenCV programs on Linux

Linux OpenCV Camera Compatibility

• For some cameras (e.g. Quickcam Express) try running "v4lctl -c /dev/video0 setinput X" to get a valid list of inputs and then set it appropriately with the same command.

MacOS X Version

OpenCV 1.0 for Mac OSX/Linux/Unix is available on the files section of SourceForge.

Until the next release of OpenCV, there is now a pre-built version (universal binary Private Framework, together with a demo application). You can obtain it from http://www.ient.rwth-aachen.de/~asbach/OpenCV-Private-Framework-1.1.dmg .

Additional information on building and on the current state of development can be found on the Mac_OS_X_OpenCV_Port page.

MacOS X Compatible Cameras

Windows Version

Windows Compatible Cameras

DirectShow without the hassle

You can use the video input library at http://muonics.net/school/spring05/videoInput/ to avoid the messiness of dealing with DirectShow. It even works for direct capture to the imageData field of an IPL image, though I found that the result was upsidedown and had red and blue switched. (Note from author of videoInput: new version out! With option to have RGB or BGR pixels and loads of other features)

Firewire 1394 cameras without Direct X

You can install the CMU 1394 driver for your camera and then use the API of this driver to capture video from the camera. In this way, you can avoid the use of DirectX. See example for details.

People

Feel free to add yourself here...

• P. Girish, India, Neomagic Semiconductors http://www.neomagic.com

• Asit Jain

• Indraneel Ahluwalia

• Gary Bradski

• Adrian Kaehler

• ChristophSeibert

• Stephen Antony

• OlivierBornet

• Parameswaran

• RalfStauder

• RicardBorras

• Adam Fourney

• PankajVallabh

• JavierBarandiaran

• JairoSanchez

• AndresCaro

• XiaomingLiu

• DominiqueBelhachemi

• mydreamcome

• Michael T. Matibag, michaelmatibag@gmail.com , micmat_aztecx@yahoo.com

• Kecsap

• Luiz Reginaldo

• Shiqi Yu

• MiguelFigueiredo

• AlexandraRibeiro

• Alejandro Reimondo

• Tiago Romagnani Silveira

• BarisEvrimDemiroz

• WernerPomwenger

• ASBG Group in Portugal

• Francisco Bernardo

• Haar Hand Tracker Losi, losteiner@gmail.com

• RomanStanchak

• VadimPisarevsky

• Le Khanh Thanh,Ha Noi, Viet Nam. comp_vis@yahoo.com

• Lismont Kevin, stvv@pandora.be , Belgium

• Thijs Triemstra

• Zhaoxiang Zhang

• ChrisFairles

• Sami Bourgou

• Shirish Goyal

• Efextra eSolutions

• Jean-Baptiste LE STANG

• Andreas Geiger - Universitaet Karlsruhe

• MoaathALRajab

• Laurent Wagneur, France, INGENICA http://www.ingenica.fr

• mailto:congthienvn@yahoo.com Pham Cong Thien, Suwon, Korea http://www.hcmuaf.edu.vn/cackhoa/kcntt/nhanvien/pcthien/index.html

• Antonio, Metron Hong Kong Ltd. http://www.metronhk.com

• Jérôme Landré, France - http://pagesperso-orange.fr/jerome.landre/index_en.xhtml

• Lee Seongjoo, Republic of Korea Artificial Intellegence Laboratory, Yonsei University

• Son Do-Lenh

• NicolasSaunier

• TamasUngi - and his tutorial notes

• Phan Anh Vu - Hanoi University of Technology - virces931511@yahoo.com - http://cntt.tv

• Oscar Deniz Suarez, Spain http://mozart.dis.ulpgc.es/Gias/oscar.html

• Bluekid

• João Quintas, Portugal

• Sunu Wibirama - OpenCV Indonesia

• Matthew Bennett - Course Leader of BTEC Games Design at Gloucester College and author of http://colourisation.net

• Rohan Anil - undergraduate at BITS Pilani Goa Campus , author of http://pam-face-authentication.googlecode.com/