基于超分辨成像算法的群目标识别研究

Target recognition based on the algorithm of super resolution imaging

超分辨成像是一种数据自适应的成像技术,可以应用于ISAR成像的群目标识别中。目前常见的Capon等算法由于最优加权矢量的计算误差,会造成波束形成器的性能急剧下降,然而超分辨成像算法加入二次约束和子空间约束后,成像效果大大提高,并对群目标识别有着很好的效果。由常见的Capon为起点,分析超分辨成像算法的二次约束和子空间约束的优势,分别比对比稳健自适应波束形成的对角加载算法和基于特征空间的波束形成算法,最后通过仿真展示结果。

Super resolution imaging is a kind of data adaptive imaging technology,which can be applied to the group target recognition of ISAR imaging.At present,the error caused by the optimal weighting vector can cause the performance of the beam forming device to drop dramatically.However,the imaging effect of the super resolution imaging algorithm is greatly improved after the addition of the quadratically constraints and subspace constraints,and has a good effect on the target recognition.Based on the Capon as the starting point,this paper analyzes the advantages of the two constraints and subspace constraints of the super resolution imaging algorithm,and compares the diagonal loading algorithm and the beam forming algorithm based on the feature space,and finally shows the results through simulation.