I am starting up a project that will require Computer Vision to be used on the Raspberry Pi. I am relatively familiar with OpenCV and am aware that it can be installed successfully on a Raspberry Pi though I don't know how successful the Pi would be with intensive graphics operations. Would an RPi be suited to basic image matching and object tracking?

As a secondary goal, we would love to be able to use OpenKinect with the Pi . . . but as far as I can tell from reading this other question and blog post. No one has been able to get video data from the Kinect.

1 Answer 1


The Pi is not powerful enough to do much. It is very hard to get webcams working at decent frame rates to do computer vision and the kinect is not possible. You can easily compile the openkinect drivers and use a powered USB hub to attach the kinect too. However there are issues with the USB bandwidth of the Pi that doesn't allow it to grab data from the kinect. Perhaps with will change in the future with newer board revisions, but not possible now.

  • So with the B+ would you say it's possible?
    – Jacksonkr
    Commented Feb 28, 2015 at 23:30
  • 1
    Try the RPi 2, it is much faster. I don't have one now, but want to try this when I do get it.
    – kevin
    Commented Mar 2, 2015 at 0:22
  • Good to know. Also, have you ever gotten openkinect compiled on your current version of rpi? I'm just curious as to what your performance was like.
    – Jacksonkr
    Commented Mar 2, 2015 at 1:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.