Research On Pulsar Signal De-noising Method Based On Wavelet Packet Threshold Method

Since the completion of the 500 m aperture fast radio telescope in Pingtang,Guizhou Province in 2016,as a powerful tool for pulsar research,fast uses a 19 beam multi beam receiver system to conduct large-scale Sky Survey and search in the sky area for more than 200 days a year,8 hours a day,search and receive the electromagnetic pulse signal from the pulsar,and determine whether there is pulsar in the sky area according to the electromagnetic pulse signal,However,due to the influence of various celestial motion and optical noise,the pulsar signal is submerged in a large amount of noise.Therefore,after the fast radio telescope receives the electromagnetic pulse signal,it must carry out operations such as de-interference,de dispersion and de noise to obtain the real pulsar signal completely,so as to find more pulsars and provide more data support for our astronomy to explore the universe.The traditional signal de-noising method is Fourier transform,while the pulsar signal is discrete and non-stationary,so the traditional Fourier transform can not play a very ideal effect.For this reason,wavelet transform de-noising method has entered the field of vision of researchers.Wavelet transform is very suitable for processing non-stationary signals because of its own multi-resolution analysis characteristics,so it is widely used in pulsar signal de-noising.However,with the further development of pulsar signal de-noising research,it is found that wavelet transform is constantly decomposing low-frequency signals every time,The high-frequency signal is not decomposed continuously,and the pulse signal which may be hidden in the high-frequency signal can not be separated,which reduces the accuracy of de-noising.Wavelet packet transform is a signal analysis method proposed to overcome the inability of wavelet transform to decompose the whole frequency band.It can remove the noise of each frequency band more accurately,and it can also decompose the high-frequency part of the hidden pulse signal.Good de-noising effect.In view of this,this paper proposes a pulsar signal de-noising method based on wavelet packet threshold method.After the original data of the pulsar signal is de-interfered and de-dispersed,the wavelet packet is first decomposed,then the threshold method is processed,and finally the pulsar signal is reconstructed.,So as to achieve the purpose of de-noising.Since the choice of wavelet basis function directly affects the experimental effect of wavelet packet threshold de-noising,the mean square error and signal-to-noise ratio are used as the selection basis in the experiment.The wavelet basis function with the smallest signal mean square and error after de-noising is selected as the wavelet packet Wavelet basis function for de-noising.This experiment uses smoothness,power spectrum,signal-to-noise ratio,peak signal-to-noise ratio,root mean square error,etc.as the de-noising effect evaluation indicators.The experimental results show that the de-noising effect of this method is significantly better than the wavelet threshold method: in the experiment Experiments were carried out on the files obtained by Parkes observation data processing and the files obtained by FAST observation data processing.The wavelet basis function and threshold value determined by wavelet transformation were selected according to the root mean square error and the signal-to-noise ratio as the selection criteria for the wavelet threshold method.De-noising and wavelet packet threshold method de-noising.The de-noising results show that the Parkes pulsar observation data adopts wavelet packet hard threshold method to de-noise the pulsar signal.The smoothness of the effect picture is 0.000068,the signal-to-noise ratio is 128.574093 d B,and the peak value The signal-to-noise ratio is 50.216641 d B,and the smoothness of the wavelet threshold method for de-noising pulsar signals is 0.000070,the signal-to-noise ratio is127.619843 d B,and the peak signal-to-noise ratio is 48.308142.The wavelet packet threshold method is smaller than the wave threshold method.The de-noising smoothness is reduced by 0.000002,the signal-to-noise ratio is improved by0.95425 d B,and the peak signal-to-noise ratio is improved by 1.908499 d B.The FAST observation data uses the wavelet packet hard threshold method to de-noise the pulsar signal.The smoothness of the effect picture is 0.000144,the signal-to-noise ratio is92.736203 d B,and the peak signal-to-noise ratio is 92.870728 d B.The wavelet threshold method de-noises the pulsar signal.The smoothness of the noise effect map is 0.000338,the signal-to-noise ratio is 86.325067,and the peak signal-to-noise ratio is 86.459591.The wavelet packet threshold method is smaller.The wave threshold method de-noising smoothness is reduced by 0.000214,and the signal-to-noise ratio is increased by 6.411136 d B.The signal-to-noise ratio has been improved by6.545661 d B,which provides a new attempt for pulsar signal de-noising.It is of great significance for pulsar signal de-noising,even observation and discovery.