Improved camshift algorithm based on weighted histogram model

In order to improve the tracking performance, an improved CamShift tracking algorithm is proposed by strengthening the distinguishing hue characteristics. The distinguishing hue of the target is determined by contrasting the target and its surrounding area. The target model is described by a weighted hue histogram. Max weights are given to the distinguishing hues to distinguish the target from the background. The probability projection image is established by the weighted histogram model. Simulation results show that the improved method has better tracking performance than the common method, and the distinguishing hue characteristics can restrain the disturbance of the background.