高频地波雷达回波信号的频谱主要由一阶峰与二阶谱构成,而这些频谱中蕴含着丰富的海况信息,因此分离其频谱具有重要的理论和现实意义。利用高频地波雷达信号的一阶峰与二阶谱信号的独立性,分别建立不同阶次的自回归(AR)模型。为了估计对应于一阶峰和二阶谱的AR 模型阶数和参数,首先利用自适应原子分裂算法对回波谱进行稀疏估计,得到一阶峰与二阶谱的混合特征根及其由特征根构成的自相关函数的系数;然后,利用特征根的组合构成两个自回归模型,分别计算对应模型的由特征根表示的自相关函数的系数,并将所得到特征根表示的自相关系数逐一进行对比,当所有的特征根表示的自相关系数都近似相等时,则实现对一阶峰和二阶谱的分离;最后,模拟数据及真实的高频地波雷达回波谱信号分离计算验证了该方法的可行性。

The spectrum of the HF ground wave radar echo signal is mainly composed of first-order peak and second-order spectrum, which contains abundant sea state information. Therefore, separating the spectrum has important theoretical and practical significance. In this paper, the independence of the HF ground wave radar signal’s first-order peak and second-order spectrum was used to establish different levels of Auto Regressive (AR) models. In order to estimate the order and parameters of the AR model which are corresponding to the first-order peak and the second-order spectrum, first, the adaptive spectrum splitting algorithm was used to sparsely estimate the echo spectrum, obtaining the mixed eigenvalues of first-order peak and second-order spectrum and the coefficients of the autocorrelation function composed of eigenvalues. Then, two autoregressive models were constructed by the combination of eigenvalues and the coefficients of the autocorrelation function represented by the eigenvalues of the corresponding model are respectively calculated. Afterwards, the autocorrelation coefficients of the obtained eigenvalues were compared one by one. When the autocorrelation coefficients represented by all the eigenvalues are approximately equal, the separation of the first-order peak and the second-order spectrum is achieved. Finally, the simulation data and the separation calculation of real HF ground wave radar echo spectrum signal verified the feasibility of the method.

TN957. 51

“图像处理与图像通信冶江苏省重点实验室资助项目(TX2015001)