论文部分内容阅读
高分辨距离像(HRRP)已成为雷达自动目标识别(RATR)领域研究的热点。由于目标复数HRRP的初相具有不确定性(本文称其为初相敏感性),在RATR中通常使用的HRRP为实向量(文中称为实HRRP),即目标复基频回波信号的幅度像,而完全丢弃了相位信息。事实上,复HRRP中除初相外的其他相位信息也应该是对识别有利的。文中提出了一种复HRRP的特征提取方法,该方法得到的复特征向量与初相无关但却保留了距离单元间的差分相位信息,称为相位差分复特征向量。若基于散射点模型理论分析,这种复特征向量的物理意义与实HRRP类似(只是包含了差分相位信息),因此,对实HRRP适用的识别、建模和预处理方法同样适用于该复特征向量。同时,相位差分复特征向量的各分量具有类Gauss性,更有利于统计识别。基于实测数据的识别实验表明,只要在特征提取时适当选取距离单元间隔参数,文中提出的相位差分复特征向量可以取得比实HRRP更好的识别效果。
High resolution range image (HRRP) has become a hot spot in radar automatic target recognition (RATR) field. Due to the uncertainty of the primary phase of the target complex HRRP, referred to herein as primary phase sensitivity, the HRRP commonly used in RATR is the real vector (referred to herein as real HRRP), the amplitude of the target complex fundamental frequency echo signal Like, but completely discarded phase information. In fact, phase information in the HRRP other than the pre-phase should also be good for identification. In this paper, a new feature extraction method for complex HRRP is proposed. The complex feature vectors obtained by this method are independent of the initial phase but retain the information of the differential phase between the distance elements, which is called phase difference complex eigenvector. If based on the scattering point model theory analysis, the complex eigenvector has the same physical meaning as the real HRRP (only contains the differential phase information). Therefore, the applicable identification, modeling and preprocessing methods for real HRRP also apply to the complex feature vector. At the same time, the components of the phase difference complex eigenvector are Gauss-like, which is more conducive to statistical recognition. The experimental results based on the measured data show that the phase difference complex eigenvectors proposed in this paper can achieve a better recognition effect than real HRRP, as long as the distance interval parameters are properly selected during feature extraction.