the source code is now in the repository:
There is only the Windows Version as an VS2008 project to compile. MAC guys will need to make their own xcode project.
There are some further changes i made to the last version:
There are now different startmodes for setting the background image at start time. Look in the conf.xml under BACKGROUND STARTUPMODE you will see what’s it all about.
You now have the possibility to save the current background image as a file for later loading at the start. (I’ve change the ‘s’ key function from saving all settings to only saving the background image.). So hit ‘s’ and look in the data folder for a new Image. The name of the Image is set with the IMAGE tag.
There is also a second feature: Image Undistorting with openCV.
Under the BOOLEAN settings i added a UNDISTORT tag where you switch on/off this function. If you use it there will be a tiny preview Image of the undistorted Image at the right bottom of the CCV window. This is only a first implementation and it’s for some CPU a performance killer. The code has all stuff to only mapping the position data of the blobs for better performance, but i deactivated it as the CCV calibration will already do some undistortion for the blob data, so there would be no further advance. A complete image undistortion make only sense for the fiducial tracking to get a better pattern recognition if you use wide angle lenses.
I made a Intrinsics.xml and a Distortion.xml for the openCV undistortion. On a second step i will publish a tool to make your own files for your specific lense. (If you can’t wait you could look at the openframeworks forum to get further information and some source stuff from other guys, or look into the openCV book from oreilly ).
CCV should be crossplatform but the Multiplexer has some Windows specific memory check method , and i had no motivation to make this crossplatform.
If there is some OpenGL expert, he should look at the FBO implementation of the Multiplexer to help finding a solution for the memory leak with the glDrawPixels call. (Perhaps i didn’t created all buffers or something like this).
feedback is welcome
I’m in contact with Christian Moore, so it’s up to him if or what features will go into the main trunk of CCV. At this time this is some sort of parallel version. (My codebase is the CCV 1.2 Version, so this is not a 1.4 Version!)