滑动窗口核岭回归运动目标轨迹预测算法  

Sliding window kernel ridge regression trajectory predicting algorithm

作　　者：赵菲[1] 卢焕章[1] 张志勇[1] 

机构地区：[1]国防科学技术大学ATR国防科技重点实验室

出　　处：《红外与激光工程》2013年第3期829-835,共7页Infrared and Laser Engineering

基　　金：国家"863"重大项目;武器装备预研基金项目

摘　　要：针对序列图像中非线性运动目标轨迹的预测问题,文中提出了一种滑动窗口核岭回归运动目标轨迹预测算法。文章推导了完整的核岭回归算法,并得出了高斯核函数的条件下滑动窗口核岭回归算法的递推形式。算法的实现基于滑动窗口的方式,在每帧图像中以最近几帧中的目标轨迹位置为输入,使用核岭回归算法对下一帧轨迹位置,进行预测。实验结果表明,该算法能够较好的预测非线性的目标运动轨迹,预测误差较小,且算法结构简单,具有较强的实用性。Due to the requirement for prediction of nonlinear target trajectory in image sequences, a sliding window kernel ridge regression（KRR） target trajectory predicting algorithm was proposed. The full KRR which posses the bias item was deduced first, and then the iterative form of sliding window KRR algorithm was also derived in this paper. The algorithm was carried out in a sliding way, and the trajectory information in latest frames was used to predict the position in the next frame, which was achieved by using the KRR. The experimental results demonstrate that the proposed algorithm can predict the nonlinear trajectories accurately, and the prediction error is small. The structure of the proposed algorithm is simple and practicable in engineering application.

关 键 词：核岭回归 滑动窗口 轨迹预测 非线性 

分 类 号：TN911.73[电子电信—通信与信息系统]