Research On Weak Signal Detection Based On HHT

The problem of weak signal detection has been the focus and focus of research in recent years both at home and abroad.Detecting weak signals early in the background of strong noise interference is of great significance to real life.Due to the fact that background noise is too large and weak signals are easily affected by noise,it is of great theoretical and practical significance to realize effective detection of weak signals in strong noise background.The HHT method is a method capable of analyzing nonlinear and non-stationary signals.It can decompose frequency-different components of a mixed signal in ascending order of frequencies to generate a series of modal components with different frequencies.The HHT spectrum and Hilbert marginal spectrum formed after Hilbert transform of these modal components can have an accurate representation of the physical process in reality.The EMD method in HHT can decompose the noise and weak signals in the noisy signal,and can further confirm whether the frequency of the weak signal to be detected exists in the noisy signal through the HHT spectrum and Hilbert's marginal spectrum.This is strong.The method of detecting weak signals under the influence of noise background provides a new direction.Therefore,this paper starts with the theoretical basis of HHT transformation.Through the research on the problems existing in HHT and effective denoising preprocessing methods,we have found a method that can detect weak signals under strong noise background,and proposed a method based on HHT.Weak signal detection new method,the main work is as follows:1)A weak signal detection method based on HHT and wavelet threshold denoising is proposed.Since the wavelet threshold denoising method has multi-resolution characteristics,this method uses wavelet threshold denoising as a noise reduction preprocessing method to reduce the influence of noise on the detection results.The endpoint effect generated by the EMD method is improved by adopting a slope-end continuation method with certain adaptability,so that the detection result is more accurate.In the strong noise background,the single-frequency signal detection has achieved good results,and good detection results have been achieved in the detection of mixed-frequency signals.2)A weak signal detection method based on HHT and digital morphological filtering is proposed.Through the research of digital morphology filtering method and EMD denoising theory,a combined denoising method based on digital morphology filtering and EMD method is proposed.The denoising effect has been greatly improved and used as a detection method.Denoising preprocessing method.Mirroring endpoint continuation method,which uses signal periodicity to perform endpoint extension on the signal,can effectively suppress the endpoint effect of the EMD method,making the detection result more accurate.In the strong noise background,the single-frequency signal detection has achieved good results,and good detection results have been achieved in the detection of mixed-frequency signals.3)A weak signal detection method based on HHT and singular value decomposition is proposed.Through the research of the singular value decomposition theory,a method for determining the dimension of the Hankel matrix and determining the effective rank order of the singular value is proposed and used as a denoising preprocessing method for the detection algorithm.Through the study of boundary-scale endpoint continuation method,this thesis has some flaws in the theory and proposes an improved method.The improved boundary-scale endpoint extension method can effectively suppress the endpoint effect of EMD method and make the detection result more effective.accurate.Through the study of the above theories,both the detection of single-frequency signals in the background of strong noise and the detection of mixed-frequency signals have achieved good results.