Create a FrameSource implementation that makes the camera image and the depth image from the Kinect available.
Added the module lecturesight-framesource-kinect. The module already contains skeleton classes
KinectVisualFrameGrabber, KinectVisualFrameGrabberFactory (FrameSource for the visual camera)
KinectDepthFrameGrabber, KinectDepthFrameGrabberFactory (FrameSource for the depth sensor)
Services should be ready to run, grabbing frames from the Kinect sensors must be implemented.
The FrameSources have the type prefix 'kinect-visual' and 'kinect-depth' so framesource MRLs will look like this:
For test purposes FrameSources now throw the signal "framesource.NEWFRAME" in addition to the normal signals that has a name that is generated at runtime. this way we can test if a framesource implementation is operational without the other components being loaded. in the configuration the following property must be set in order to make the heart beat listen only to framesources signaling that a new frame was acquired:
Kinect FrameSource MRLs just contain the device index since libfreenect does not provide device nodes:
The implementation also contains a pass-thru for libfreenect's logging. It can be activated with the following VM property
freenect.log = (true|false)
and the log level for the libfreenect logging can be set with VM property
freenect.log.level = (FATAL|ERROR|WARNING|NOTICE|INFO|DEBUG|SPEW|FLOOD)
A typical logging configuration would be