基于视轴矢量序列的目标轨迹确认方法  

Target Confirmation Method Based on Line of Sight Vector Sequence

作　　者：薛永宏[1] 王铁兵 乔凯[1] 樊士伟[1] XUE Yonghong;WANG Tiebing;QIAO Kai;FAN Shiwei(Beijing Institute of Tracking and Telecommunication Technology,Beijing 100094,China)

机构地区：[1]北京跟踪与通信技术研究所,北京100094

出　　处：《红外技术》2022年第4期357-363,共7页Infrared Technology

摘　　要：针对椭圆轨道红外监视系统图像畸变大,基于图像配准进行目标轨迹确认误差大的问题,提出基于视轴矢量序列的目标确认分析方法;建立了目标视轴矢量序列的状态转移模型和观测模型,利用GM-PHD(Gaussian mixture probability hypothesis density)滤波器对目标视轴矢量进行滤波实现目标确认分析;针对在目标状态未知条件下,GM-PHD滤波器易于丢失新目标的问题,采用反馈滤波方式对GM-PHD滤波器进行改进设计。仿真结果验证了基于视轴矢量序列进行目标轨迹确认分析的有效性以及改进GM-PHD滤波器对新目标状态估计的时效性。Aiming at the serious image distortion of elliptical orbit infrared surveillance system and the low precision of target confirmation algorithm based on image registration,a novel target confirmation method based on line of sight(LOS)vector sequence is proposed.By establishing the state transition model and observation model of target LOS vector sequence,the GM-PHD filter is used to filter LOS vector sequence data and confirm target.Owing to that the GM-PHD filter will lose the new target when its states are unknown,the feedback GM-PHD filter is proposed.The simulation results verify the effectiveness of the proposed method based on LOS vector sequence and the timeliness of the improved GM-PHD feedback filter for new target states estimation.

关 键 词：目标轨迹确认 视轴矢量序列 天基红外监视系统 GM-PHD滤波器 随机有限集 

分 类 号：TP911.73[自动化与计算机技术]