一种新的红外弱小目标检测前跟踪算法

Novel track-before-detect algorithm for small infrared target

针对低信噪比复杂环境下红外弱小目标检测问题,提出了一种基于拟蒙特卡罗抽样的检测前跟踪算法.采用新的基于拟蒙特卡罗抽样的高斯粒子滤波算法(QMC-GPF)估计目标运动状态和红外小目标幅度状态,同时利用该滤波算法中迭代更新的协方差矩阵的收敛特性构建判断逻辑,实现目标"软判决"检测.对仿真红外图像序列的实验表明,该算法可快速有效地跟踪和检测信噪比不小于1dB的弱小目标.

A novel track-before-detect filtering algorithm is proposed for the dim infrared target with a low signal-to-noise ratio in complex backgrounds. A new particle filter called the Quasi-Monte Carlo sampling based Gaussian particle filter（QMC-GPF） is developed to estimate on-line the standard kinematic parameters of the target, including the position and velocity, as well as the amplitude of the target. The convergence characteristic of the covariance of the posterior densities propagated in the QMC-GPF is used to construct the logical rules for soft-decision detection of a possible target. The algorithm is tested with a synthetic target in IR image sequences, and it is proved that the algorithm is capable of performing sufficiently well for the dim target of SNR≥1 dB.