This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
One option for a video feed would be to add a Kalman filter. This can prevent brief periods where you stop detecting the object, and it works even if the object becomes occluded. A kalman filter will compare the previous state of an object to the newly provided state to determine if there is an object and where it is. In laymens terms it does something like this: well I've been seeing two objects for the past 10 seconds and now ones gone, whats the probability that one warped away (probably not so high). I will continue tracking it and predict where I think it is. --5 seconds pass-- Is the object still gone, yes, okay now I will stop tracking it.
