1 | initial version |
it's probably only a matter of supplying the right constant to the VideoCapture constructor, or the open method of that.
in c++, it works like this:
VideoCapture cap(0); // takes 1st cam. on win, that would be using vfw, the default
VideoCapture cap(CV_CAP_DSHOW + 0); // takes 1st cam but tries to use the directshow path
VideoCapture cap(CV_CAP_OPENNI_ASUS + 1); // get the 2nd asus depth sensor, for a change ;)
// lacking hardware, i can't test, but i bet cmu goes like this:
VideoCapture cap(CV_CAP_CMU1394 + 0); // CV_CAP_FIREWARE, CV_CAP_IEEE1394 should work, too, boils down to the same enum.
those consts are in highgui_c.h, l 279 and down. i'm unsure how to access them from python, maybe its cv2.cv.CV_CAP_something, but no idea here
2 | No.2 Revision |
apart from the question if cmu support was builtin with your current libraries, it's probably only a matter of supplying the right constant to the VideoCapture constructor, or the open method of that.
in c++, it works like this:
VideoCapture cap(0); // takes 1st cam. on win, that would be using vfw, the default
VideoCapture cap(CV_CAP_DSHOW + 0); // takes 1st cam but tries to use the directshow path
VideoCapture cap(CV_CAP_OPENNI_ASUS + 1); // get the 2nd asus depth sensor, for a change ;)
// lacking hardware, i can't test, but i bet cmu goes like this:
VideoCapture cap(CV_CAP_CMU1394 + 0); // CV_CAP_FIREWARE, CV_CAP_IEEE1394 should work, too, boils down to the same enum.
those consts are in highgui_c.h, l 279 and down. i'm unsure how to access them from python, maybe its cv2.cv.CV_CAP_something, but no idea here