Ask Your Question

Passing "hints" to Object Tracking algorithm?

asked 2016-04-11 06:20:00 -0500

dtbaker gravatar image

Is it possible to pass "hints" to any of the object tracking algorithms to improve performance?

Having a robot with onboard object tracking, I know my robot is about to turn left so I wonder if I can tell the "object tracker" that it should expect the tracked object to appear at position x/y in the next video frame? And if this could improve any processing speed?

(I'm running on a low powered Raspberry Pi so trying to squeeze as much out of it as possible.)

Examples of movement are:

  • I am tracking an object straight ahead of me but I have to do a slight detour around this obstacle infront of me first. This sharp detour could see the tracked object move from the center of the camera all the way to the edge of the camera in the very next frame. I could tell it that we're turning left so try to find the object on the right of the next video frames.
  • There is also some raw vectors coming straight off the GPU which might be able to used as "hints" towards the more CPU intensive object tracker

Any thoughts?


edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2016-04-11 06:50:50 -0500

What you intend to use is a prediction/correction approach. Fortunately, OpenCV has one built-in: the Kalman filter. Have a look at the documentation and the sample in cpp/kalman.cpp. I also suggest you have a look at the Wikipedia page (as referenced in the documentation) to be sure you understand how it works).

Basically, the Kalman filter will make a prediction of where the next position should be, that you will correct with the measurement you retrieve (you robot next move probably in this case). You need also to model the noise of the observation. The "Learning OpenCV" book has a section about this prediction/correction process (not sure about the upcoming release, but it should have it also).

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower


Asked: 2016-04-11 06:20:00 -0500

Seen: 95 times

Last updated: Apr 11 '16