粒子滤波和均值漂移相结合目标跟踪算法

采用粒子滤波算法解决运动目标跟踪中非线性非高斯问题。将均值漂移算法嵌入到粒子滤波的采样阶段中,通过将每个粒子聚集到所在区域的局部极值,提高了采样粒子的使用效率。当发生目标遮挡时采用改进的粒子滤波算法,当无遮挡时采用均值漂移算法以提高速度。实验结果表明,该方法较传统单一算法具有较强的实时性和鲁棒性,能够有效实现在遮挡场景下的目标跟踪。

Object Tracking Based on Particle Filter and Mean-shift

The article presents the integration of mean-shift algorithm with particle filtering during the object tracking.In our method mean-shift algorithm is used in the sampling steps of particle filtering,which efficiently reduces the number of sampled particles.Under occlusion the improved particle filter algorithm is employed while under non-occlusion mean-shift algorithm is adopted.When applied in the moving object tracking,our method proved to be more robust and time saving compared with the conventional part...