I am using the 5M camera module to capture images and using the Leopard Board as a USB camera currently. I want to do some edge detection on the images before I stream them. Can anyone tell the best way to do this?
Thanks an regards
I wanted to do something comparable and I spent a lot of time looking at how I could manipulate gstreamer to allow for me to run a video processing/computer vision algorithm on the streaming video.
After a lot of effort I gave up trying to use gstreamer for video processing and interacted directly with the V4L2 driver provided.
Information on the V4L2 driver can be found here: http://v4l2spec.bytesex.org/spec-single/v4l2.html (appendix b contains a "frame grab" example).
Hope that helps. Has anyone else used gstreamer for computer vision algorithms successfully?