Research of Object Tracking Algorithm Based on Machine Vision
In order to solve the problem of tracking drift in occlusion,an object tracking algorithm(MOSKal)based on MOSSE and Kalman filter is proposed to improve the tracking effect.The algorithm uses the correlation measurement between the object and the MOSSE filter to judge whether there is occlusion or not in every frame.When there is no occlusion,the MOSSE filter is used as the main tracker,the MOSSE filter is updated and the Kalman filter is modified according to the obtained position;when there is occlusion,the Kalman filter is set as the main tracker to predict position,and the MOSSE filter is kept unchanged.Qualitative and quantitative analysis of experiments were made from the aspects of tracking speed,accuracy and success rate.The results show that the algorithm can track target quickly and effectively with or without occlusion.Compared with other similar algorithms,the algorithm can effectively improve the tracking effect under partial and complete occlusion,and has strong robustness.
国家科技期刊平台
NSTL
万方数据
