基于粒子滤波改进的VTS微弱目标检测前跟踪算法

舰船交通服务系统是民用雷达的信息集成系统,探测微弱目标存在RCS小、回波弱、杂波强等问题,导致信噪比低,难以实现有效检测跟踪。基于粒子滤波的检测前跟踪技术对低信噪比下微弱目标信息积累和探测有良好效果。通过采集单设备实测数据,构建遗忘因子和收敛因子以增加重采样的效率,引入虚拟采样保持粒子的多样性,提升粒子滤波对微弱目标的探测能力。仿真试验表明,改进后的算法可实现舰船交通服务系统对微弱目标的有效探测,并能获得较精准的目标状态估计值。

An Improved PF-TBD Algorithm for VTS System Dim Objects Detection

VTS (Vessel Traffic Services)system is a very important civil Radar surveillance furnishment,which is challenged by various problem especially when monitoring objects with small RCS in a low SNR or strong noise scene; Track before detection (TBD)algorithm based upon Particle filter (PF)takes advantage for its adaption in solving non-linear,non-Gaussian or unsteady state problems, especially when detecting and tracking dim object in a low SNR scene. In this paper synthetic sample strategy,fading factor and convergence factor are integrated to improve sampling performance and keep diversity of the particle. Analysis and experiment proves that this promoted algorithm can detect and track faint target in VTS radar video and approximate target’s status more precisely compared with original EPF-TBD algorithm at cost of a few more computing burden.