Forums > Jitter

Tracking radial motion using cv.jit


d_s
February 7, 2007 | 5:47 pm

Hello All,

I am interested in tracking radial motion in a room, that is whether people are moving clockwise or counter clockwise related to a fixed point on the floor. I am using an overhead camera.

I’ve looked at the optical flow patches in cv.jit but can’t quite figure out how to derive radial motion from these.

Would tracking blob centroids be better and look for vectors?

Any advice is appreciated.

Best,
d_s


February 7, 2007 | 6:10 pm

Damon Seeley skrev:
> Hello All,
>
> I am interested in tracking radial motion in a room, that is whether people are moving clockwise or counter clockwise related to a fixed point on the floor. I am using an overhead camera.
>
> I’ve looked at the optical flow patches in cv.jit but can’t quite figure out how to derive radial motion from these.
>
> Would tracking blob centroids be better and look for vectors?
>
I’m no tracking wiz, but jit.cv.blobs.centroids (or whatever it’s
called) outputs x/y coordinates for the blobs, right? I would send
these through a [cartopol] and then sample the theta output. There’s a
bit of trickery involved while crossing from 360 to 1 degree as far as I
remember, but that can be solved with some logic.

Andreas.


Viewing 2 posts - 1 through 2 (of 2 total)