robb-broome/Install opencv OSX Mavericks.md

## Install opencv OSX Mavericks.md

      
    Raw
  

              Install opencv OSX Mavericks.md
            
          
    #Install opencv MacOSX Mavericks

fire up xcode - to insure that command line tools are there.
xcode-select --install (older OS?)
https://developer.apple.com/downloads/index.action# (you need an apple developer account, use your apple id)
brew update
brew tap homebrew/science

Cloning into '/usr/local/Library/Taps/homebrew-science'...
remote: Reusing existing pack: 3765, done.
remote: Counting objects: 63, done.
remote: Compressing objects: 100% (63/63), done.
remote: Total 3828 (delta 27), reused 0 (delta 0)
Receiving objects: 100% (3828/3828), 974.71 KiB | 1.38 MiB/s, done.
Resolving deltas: 100% (1782/1782), done.
Tapped 244 formula

If your py is in good shape you may be able to skip down to 'brew install opencv'. No harm in trying.

brew uninstall python
brew unlink python
brew install python - (many bad links to delete here, ended up forcing it)
  	> brew link --overwrite python
 	Linking /usr/local/Cellar/python/2.7.6... 35 symlinks created
 

pip install numpy...  expect a ton of very helpful warnings and complicated-looking output, followed by
  Successfully installed numpy
 Cleaning up...
 

Add to .bashrc or .zshrc:

export PYTHONPATH=/usr/local/Cellar/opencv/2.4.7.1/lib/python2.7/site-packages:$PYTHONPATH

finally, in a new terminal window:
 brew install opencv
➜  we-main git:(feature/112-org-thumb) ✗ rvm:(ruby-2.0.0-p247@we) brew install opencv
==> Installing dependencies for opencv: cmake, libpng
==> Installing opencv dependency: cmake
==> Downloading https://downloads.sf.net/project/machomebrew/Bottles/cmake-2.8.12.2.mavericks.bottle.2.tar.gz
######################################################################## 100.0%
==> Pouring cmake-2.8.12.2.mavericks.bottle.2.tar.gz
 /usr/local/Cellar/cmake/2.8.12.2: 701 files, 36M
==> Installing opencv dependency: libpng
==> Downloading https://downloads.sf.net/project/machomebrew/Bottles/libpng-1.5.17.mavericks.bottle.1.tar.gz
######################################################################## 100.0%
==> Pouring libpng-1.5.17.mavericks.bottle.1.tar.gz
 /usr/local/Cellar/libpng/1.5.17: 15 files, 1.0M
==> Installing opencv
==> Downloading https://github.com/Itseez/opencv/archive/2.4.7.1.tar.gz
######################################################################## 100.0%
==> Patching
patching file cmake/OpenCVFindOpenNI.cmake
==> cmake .. -DCMAKE_INSTALL_PREFIX=/usr/local/Cellar/opencv/2.4.7.1 -DCMAKE_BUILD_TYPE=None -DCMAKE_FIND_FRAMEWORK=LAST -DCMAKE_VERBOSE_MAKEFILE=ON -Wno-dev -DCMAKE_OSX_DEPLOYMENT_TA
==> make
==> make install
==> Caveats
Set PYTHONPATH if you need Python to find the installed site-packages:
  export PYTHONPATH=/usr/local/lib/python2.7/site-packages:$PYTHONPATH
==> Summary
/usr/local/Cellar/opencv/2.4.7.1: 217 files, 39M, built in 113 seconds


Get a look
brew info opencv

opencv: stable 2.4.7.1
http://opencv.org/
/usr/local/Cellar/opencv/2.4.7.1 (217 files, 39M) *
  Built from source
From: https://github.com/homebrew/homebrew-science/commits/master/opencv.rb
==> Dependencies
Build: cmake ✔, pkg-config ✔
Required: libpng ✔
Optional: eigen ✘, libtiff ✔, jasper ✔, tbb ✘, qt ✔, openni ✘, ffmpeg ✘
==> Options
--32-bit
	Build 32-bit only
--c++11
	Build using C++11 mode
--with-eigen
	Build with eigen support
--with-ffmpeg
	Build with ffmpeg support
--with-jasper
	Build with jasper support
--with-libtiff
	Build with libtiff support
--with-openni
	Build with openni support
--with-qt
	Build the Qt4 backend to HighGUI
--with-tbb
	Enable parallel code in OpenCV using Intel TBB
--without-opencl
	Disable gpu code in OpenCV using OpenCL
==> Caveats
Set PYTHONPATH if you need Python to find the installed site-packages:
  export PYTHONPATH=/usr/local/lib/python2.7/site-packages:$PYTHONPATH

##Notes on implementation

example using opencv to detect a face: https://github.com/ruby-opencv/ruby-opencv/blob/master/test/samples/haarcascade_frontalface_alt.xml.gz
Get the points from opencv: http://www.ruby-doc.org/gems/docs/o/opencv-0.0.6/OpenCV/CvPoint.html