Research on application of wavelet analysis in signal processing of extremely low frequency magnetic targets

Jun Li,Bin Tian,Yang Li,Linjie Yao
DOI: https://doi.org/10.1117/12.2651864
2022-01-01
Abstract:Based on the measured signals of extremely low frequency underwater magnetic targets, this paper extracts the magnetic anomaly data of sensors at 50Hz for denoising processing, and uses the denoised magnetic anomaly data as the basic data. Random noise is added to the basic data as a simulated noise signal. Based on Stein's unbiased likelihood estimation threshold method, two wavelet family Symlets, Daubechie and soft and hard threshold selection methods commonly used in denoising literature are used to denoise the measured underwater magnetic target signal data. The different de-noising results are calculated and compared. The de-noising signal-to-noise ratio of the data is used as the evaluation index, and the de-noising effect and reliability of different combined methods are objectively analyzed. Research shows that the hard-thresholding of sym7 as the generating wave function has the lowest signal-to-noise ratio and poor denoising effect. The soft threshold processing using db9 as the generating wave function has the highest signal-to-noise ratio, which can achieve the best denoising effect.
What problem does this paper attempt to address?