由于瞳孔目标运动的非线性非高斯性,但是粒子滤波(Kalman Filter)瞳孔跟踪算法能够很好地解决非线性非高斯问题,卡尔曼滤波(Particle Filter)算法是对瞳孔目标状态的最优估计.将粒子滤波算法和卡尔曼滤波算法组合可以获得更为精确的瞳孔跟踪结果.粒子滤波的结果作为卡尔曼滤波的观测值,并且根据粒子滤波的跟踪结果确定卡尔曼滤波的转换矩阵,最终达到理想的瞳孔目标跟踪性能.%Due to the nonlinear motion of non target pupil Gauss(Kalman Filter),but the particle filter tracking algorithm can solve the nonlinear pupil good non Gauss,Calman filter(Particle Filter)algorithm is the optimal state estimation of target pupil. The results show that the combination of particle filter and Calman filter can get more accurate results.As a result of the Calman fil-ter,the particle filter is used to determine the transformation matrix of the Calman filter according to the tracking results of the parti-cle filter.
展开▼
