The FrameGrabM toolkit provides a quick way to grab still frames from a live video source, such as a FireWire video camera or a USB webcam. Individual video frames are returned to MATLAB as a matrix of RGB values. The toolkit uses LTI-CIVIL, a Java-based library instigated by Ken Larson for capturing images (see http://lti-civil.org/). Please refer to the README file for instructions on how to download and install LTI-CIVIL.
Key features of FrameGrabM:
- Targeted for Windows, Mac OS X, and Linux x86 and x64
- No need for MATLAB Image Acquisition Toolbox
- Simultaneous capture from multiple video devices
It's easy! Here is an example MATLAB script for capturing a video frame from the default capture device:
fgrabm_start % Initialize the framework and start capture
pause(2) % Wait for the capture device to warm up
pixels = fgrabm_grab(); % Grab the most recent frame and return data
image(pixels) % Display the pixels
fgrabm_shutdown % Uninitialize the capture framework
Please look at the README file for details on installation, as MATLAB's Java library path in its librarypath.txt and classpath.txt files must be configured to refer to included LTI-Civil native library files. Depending upon how MATLAB was installed, you may need administrative/root privileges to successfully perform these changes. README also contains special information pertaining to the Mac OS X platform.
Note that this is a preliminary release and I welcome feedback that can help lead to a final version. See README for more details.