传统炮位雷达在利用步进频信号进行一维距离像宽带合成时,存在回波数据量大、信噪比低等问题,系统复杂度高且成像质量不高,不利于弹丸目标的快速分类识别。针对上述问题,文中充分发掘并利用步进频信号的回波特性,将压缩感知思想应用其中,构建了基于多脉冲回波信号的广义联合块稀疏模型,提出了模型下的重构算法,并将字典学习算法与之结合,大大提高了低测量值、低信噪比情况下的弹丸目标一维距离像质量,降低后端数据处理量的同时提高了炮位雷达目标识别的正确率。理论分析和仿真实验均证明了所提算法的有效性。

When traditional weapon positioning radar uses stepped frequency signals for wideband synthesis of one-dimensional range profiles, there are problems such as large echo data and low signal-to-noise ratio. The system has high complexity and low imaging quality, which is not conducive to the rapid classification of projectile targets recognition. In response to the above problems, this paper fully explores and utilizes the echo characteristics of stepped frequency signals, apply the idea of compression perception, constructs a generalized joint block sparse model based on multi-pulse echo signals, proposes a reconstruction algorithm under this model, and combines the dictionary learning algorithm is combined with it to further improve the imaging quality of projectile targets under the conditions of low measurement values and low signal-to-noise ratio, reduce the amount of back-end data processing, and improve the accuracy of the gun-position radar target recognition. Both theoretical analysis and simulation experiment prove the effectiveness of the proposed algorithm.

TN957