Target tracking using predicted Camshift
Bo Yang, Hongjun Zhou, Xue Wang · 2008
Camshift used color histogram as a characteristic, and used mean shift as the search algorithm. The direction of grads ascension was used to reduce the characteristic match time, so that the object orientation could be faster. Camshift could accurately follow the moving target under the simple background, but at a complicated one it was very difficult for Camshift to accurately track a target. Predicted Camshift was used to track the real-time moving target in this paper. It takes the centroid of the Camshift as the observation signal, and adopts the Kalman filter to predict and correct the targetpsilas centroid. With this algorithm, it can accurately track the real-time moving target even under the interferential background.