利用PSO估算Lucas-Kanade光流模型的参数

Estimation of Lucas-Kanade Parameters Using PSO

在使用Lucas-Kanade光流法进行目标跟踪时,由于目标本身存在旋转、位移、缩放等情况,导致估计参数偏差大而影响跟踪的准确性.因此提出使用PSO对Lucas-Kanade光流法中参数做最优化处理,估算出有效参数范围,以取代传统的区域求解法.实验结果表明,该算法能快速有效地跟踪目标.

When Lucas-Kanade Algorithm was used for tracking, the deviation area of estimation parame- ter increased, due to rotation, displacement, and scaling of the target itself, and the tracking accuracy was weakened. A new algorithm was proposed by the Lucas-Kanade parameter optimization using PSO, and computed a suitable range of parameters instead of local solutions by the traditional method. Simula- tion results showed thnt thi~ nlm^Titl^m 1A ~l^kl.. ~.~ ：_~ _t.：