基于全相位滤波的密集成分的频率估计

为提高密集谱成分的频率估计精度，本文提出基于全相位滤波的频谱校正法。首先，构建了可以反映以谱峰为中心的3根谱线的相位线性偏离因子，用以辨识所关注的频点是否存在密集谱分布；其次，将所关注的频点参数代入全相位窄带滤波器的解析公式，生成滤波器系数；最后，用现有的谱校正法(如比值法)对全相位滤波输出的有效样本段进行频率估计，即可估测出密集频率成分。实验证明，本文方法可以大大地去除谱间干扰误差，相比于现有的频谱校正法，提升了密集频率成分的估计精度，因而具有较广泛的应用前景。
Abstract：To enhance the frequency estimation accuracy for densely-intervaled components,this paper proposed an all-phase filtering based spectrum correction method.Firstly，a factor reflecting the phase linearity deviation of 3 spectral lines centered with the spectral peak was constructed,which can identify whether dense components exist around the frequency position of interest.Secondly,substituting this frequency component into the analytic formula of all-phase narrow-band filter yields all the filter tap coefficients.Lastly,the estimation results were acquired by implementing an existing spectrum correction operation (such as the ratio correction) on the effective sample segment of the all-phase filter’s output.Numerical results show that,the proposed estimator can greatly remove the inter-spectra interference errors and thus achieves a higher accuracy than the existing spectrum method,which possesses a vast potential of applied prospects.