基于稀疏重构的杂波环境下红外图像空间邻近目标超分辨方法

基于稀疏重构的超分辨方法是应对空间邻近目标的有效方法之一，但是当目标处于杂波环境下时，杂波会布满在整个视场范围内，导致场景原有的稀疏性被破坏。针对这一现象提出了一种在杂波环境下的超分辨方法。该方法充分利用了传感器的结构特性以及重构算法中的参数，通过建立观测信号的红外成像模型并利用像元网格划分的方式，建立空间邻近目标群的位置和幅度信号的稀疏表示，并利用其光学系统的点扩散函数来构造超完备字典，最后通过控制重构场景中非零元素的个数比例来使重构参数处于一个合理的区间范围，以此来达到去除杂波干扰并准确重建稀疏目标的目的。

The super resolution method of infrared image closely spaced objects based on sparse reconstruction in a single frame space

The super resolution method based on sparse reconstruction in a single frame space is one of the effective method to process closely spaced objects image ,but when objects are in clutter environment ,the clutter will fill the scope of the entire field of the view to damage the sparse feature of the objects .Aiming at the phenomenon ,a clutter environment of super resolution method is put forward .The structure characteristic of the sensor and the parameters in the reconstruction algorithm are fully used in the method .Through setting up the measurement of infrared imaging model and using meshing generation ,the space representation of the location and amplitude signal of sparse adjacent to the objects group is set up ,and a overcomplete dictionary is constructed by the point spread function of the optical system . Finally , the reconstruction parameters are rebuilt by controlling the proportion of the number of elements to make them in a reasonable range ,and so the goal of removing clutter interference is achieved .