一种空间邻近目标红外多传感器立体分辨方法

空间邻近目标在红外像平面的成像因相互交叠而形成簇状像斑,对红外传感器的信号处理提出了分辨的要求。为实现对空间邻近目标的立体跟踪定位,提出基于量子粒子群优化的空间邻近目标红外多传感器立体分辨方法。在对目标像平面成像建模基础上,构建基于最小二乘准则的空间邻近目标立体分辨目标函数,针对目标函数的高维非线性特点,以QP-SO解决目标函数优化问题,估计目标空间位置。仿真结果表明:相比于传统的先单传感器像平面分辨后多传感器视线交叉定位方法,此法具有更优的目标位置估计精度、辐射强度估计精度和目标个数估计正确率。

An algorithm of stereo resolution for closely spaced objects via infrared multi-sensor

The closely spaced objects （CSO） create blur pixel-clutter on the infrared focal plane, raising anew demand for the resolution of CSO. For the stereo spatial tracking and location of CSO, a method of stereo spatial resolution for CSO based on quantum-behaved particle swarm optimization （QPSO） is presented. The method first models the object image on the infrared focal plane, then constructs a least-square stereo resolution objective functiom subsequently, QPSO is adopted here to cope with the optimization problem of the objective function arising from its high dimensionality and nonlinearity, ultimately realizing the stereo resolution of CSO. Simulation results show that, in contrast to the traditional method of single sensor resolution plus stereo location from multi-sensor, the proposed method gains a better correct probability of object number detection, as well as a better precision of object location and intensity estimation.