基于峰值特性判定模型更新的鲁棒视觉跟踪算法OA

In order to solve the problem that the traditional model updating algorithm has poor robustness in the case of occlusion,illumination change and self-rotation in visual tracking,this paper proposes a robust visual tracking algorithm of using peak characteristics to selectively update the model.In this algorithm,the target posi-tion is first determined through particle filtering tracking,then the current model is used to conduct a local exhaus-tive search near the result location of the current frame tracking,and the detected peak distribution is also used to determine the numerical matrix of target confidence.Finally,the peak-to-sidelobe ratio threshold judgment meth-od is employed to decide whether or not to update the current model.Simulation results show that the proposed al-gorithm can effectively update the target model,and that compared with the contrast algorithm,it can achieve a better tracking effect on the whole in dealing with the situation of usual occlusion,illumination change,self-rota-tion,etc.in visual tracking.