基于卡尔曼滤波的台球跟踪技术研究

Research on Tracking Technique Based on Calman Filter Pool

针对无遮挡情况下的跟踪问题,利用卡尔曼滤波的预测性与特征匹配相结合的方法对目标台球进行跟踪,以期提高跟踪的实时性和准确性。针对有遮挡情况下的跟踪问题,根据运动台球具有明显的颜色特性,采用卡尔曼滤波与Mean Shift相结合的方法,实验结果表明该方法具有良好的稳健性和实时性。

For no occlusion tracking problem under the condition of using the method of characteristics, prediction and Calman if lter matching to track the target billiards, in order to improve the real-time and accurate tracking. Aiming at the problems with occlusion tracking case, according to the motion of billiard has obvious characteristics of the color, method using the Calman if lter and Mean Shift combination, the experimental results show that this method has good robustness and real-time.