I have positions of an object moving around an image. I think I'm detecting it the best I can where most of the time I'm detecting the center of the object. However, I'm still getting the odd detection of around the center caused by the frame rate not being fast enough, and the frame containing two positions of the object.
As I can't control the frame rate, how can I minimise the effects of the noise in the jittery positions.
As this is a common issue in computer vision, are there any filters in opencv to deal with noisy position data?
I asked for the comment by Berak to be made into an answer, but my request was "declined". So, yes, the answer that I found most useful was to use the Kalman filter which is implemented in opencv.