基于双谱滤波的脉冲星信号消噪方法

Denoising pulsar signals by bispectrum filters

提出了一种基于双谱相位加幅值滤波的脉冲星信号消噪方法.首先将含噪脉冲星信号变换到双谱域,利用α-删减相位滤波器对其双谱相位进行平滑滤波,并且设计了幅值α-删减加权中值平均滤波器对其双谱幅值进行滤波处理,然后再根据滤波后的双谱值来重构信号的功率谱和相位.仿真结果表明,重构的信号不仅可以消除高斯噪声,而且还对非高斯噪声进行了有效地滤波,比直接双谱重构和仅对双谱相位进行处理重构的效果都要好.

A method to denoise pulsar signals is proposed by bispectrum phase and amplitude filtering. The noisy pulsar signal was transformed into bispectral domain; the bispectrum phase was filtered by α-trimmed filter, and the amplitude α-trimmed filter was introduced to process bispectrum amplitude. Then, the pulsar signal can be recovered from the filtered bispectrum. The results indicate that the reconstruction signal can eliminate Gaussian noise and filter non-Gaussian noise effectively. This method result is much better than effects of direct bispectrum reconstruction and only dealing with spectral phase.