Ask Your Question

Calculating motion/movement energy/difference

asked 2014-04-29 16:25:27 -0500

matzahn gravatar image

updated 2015-10-06 13:26:17 -0500

Hi everybody, I am completely new in openCV and have not that much background in computer vision at all. What I need to achieve is a calculation of the "movement energy" in a real time video stream, let's say provided by depth raw data coming from a kinect camera. Concrete application: I want to measure und visualize on a scale how much/strong a large group of people, which is covered by a kinect camera, is moving. Since I found no approaches of the side of the kinect sdk and instantly by google I would like to know if there are any provided functionalities that I can use to reach this goal? I am really happy about suggestions and tips. Thanks a lot in advance.

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2014-04-29 16:58:36 -0500

adrians gravatar image


It seems that you want to compute the optical flow on an image.

I don't know if the depth data from the kinect can be integrated in the computation, but the 2d color images should be enough.

Have some references

edit flag offensive delete link more


Thanks a lot for this approach! I guess this could be the right way. But my problem is that I think it will be necessary to integrate depth data since I want to use the system in a club context with heavy light changes in the environment. Am I right that there must be always constant light conditions when using optical flow?

matzahn gravatar imagematzahn ( 2014-04-29 18:01:57 -0500 )edit

After reading a bit more about kinect integration, I see that you can get a matrix which contains depths (one channel) instead of colors.

You can try to apply an optical flow algorithm on a depth-matrix instead of a colour-matrix. (and maybe add some image filtering - smoothing, eroding, to remove the noise/background elements)

adrians gravatar imageadrians ( 2014-04-30 03:14:41 -0500 )edit

Question Tools


Asked: 2014-04-29 16:25:27 -0500

Seen: 664 times

Last updated: Apr 29 '14