It's my first time developing on OS X, so I thought I'd document the process to get started using vision algorithms on a Mac.
Platform: OS X Version 10.5.8
1. Install XCode from the Mac OS X install CD.
2. Download and install MacPorts, a command-line tool to install software and deal with dependencies for you. (For some reason I received a "Could not find specified message for index 16" error when trying to install the version for Snow Leopard, so I installed the version for Leopard instead.)
Don't forget to run sudo port selfupdate once it's installed.
3. Install OpenCV by running the command sudo port install open cv . This installs OpenCV 1.0, which is fine for me, because that's what the Python bindings (next) are tested on.
4. Download ctypes-opencv source and demo files. Once that's done, navigate to the ctypes-opencv directory and install using sudo python setup.py install
5. Open your ~/.profile (created when installing MacPorts) and add the following line:
Save and close it, then restart your terminal window for the .profile to be loaded.
6. Try one of the ctypes-opencv demos to check if it works: python houghlines.py
Running the camshift demo, using the integrated MacBook iSight webcam:
That's me thinking "Oh, OpenCV installation. Can start my research now?"