基于矩阵束的多波段雷达信号高精度融合成像算法

多波段信号融合技术在信号层将多个不同子带融合成一个大带宽信号,因而能够有效提高雷达图像距离分辨率。目前，基于全极点模型的融合技术主要采用root-MUSIC(multiple signal classification)及其改进算法实现极点的估计,在较弱的噪声条件下这种方法得到了不错的融合效果。然而在低信噪比条件下root-MUSIC算法容易受到噪声干扰而难以实现正确极点获取,进而极大影响到最终信号融合效果。为减小噪声影响,提出用矩阵束算法实现多波段信号极点估计,在此基础上通过不同子带对应极点间的相位关系估计出相干参数,同时对融合结果以信号差的2范数最小为准则进行迭代,以减小融合信号的误差。最后采用加权寻优的方式进一步提高了信号的融合精度。仿真实验结果表明，提出的方法有效提高了低信噪比条件下的多波段信号融合效果。

Multiband radar signal high precision fusion imaging method based on matrix pencil algorithm

The multiband signal fusion technique can improve the range resolution of the radar image efficiently by fusing several subbands to a great wide band signal in the signal level. At present, the all-pole model based fusion algorithms mainly apply the root multiple signal classification (MUSIC) and its modified method to estimate the signal poles, and good results are obtained in weak noise conditions. However, the root-MUSIC method which is fragile in noise could not easily get proper poles in low signal-to-noise ratio(SNR). In order to eliminate the influence of noise, a matrix pencil algorithm based method is proposed to estimate the multiband signal poles. And then the coherent parameters of multiband signals are calculated through the relation of corresponding poles. At the same time, an iterative algorithm which aims to minimize the 2-norm of signal difference is introduced to reduce signal fusion error. Apart from that a weighted optimization method is also presented to improve the fusion precision. Applications to simulate dada verify that the proposed method could get better fusion results at low SNR.