Feature Extraction And Analysis Of Underwater Acoustic Signal With Laser Interferometry

In this paper,the laser interferometer is used to design the optical path and build the experimental platform.The experimental study of underwater acoustic signal extraction is carried out to verify the feasibility of laser coherent detection of underwater acoustic source.The natural water surface fluctuation is extracted,and the data of three groups of different positions,different frequencies and different amplitudes are compared and analyzed in this study.Moreover,Filtering,short-time Fourier transform,wavelet analysis,Wigner-Ville distribution and time-frequency reassignment are used in order to achieve signal de-noising and feature signal enhancement on the basis of the traditional spectrum analysis.Experimental results show that:(1)Water natural fluctuations are mainly in the low frequency region below 500 Hz,and the water wave frequency range can be determined as a reference for further filtering and time-frequency analysis,which makes it easier to distinguish between natural fluctuation frequency and signal frequency.(2)Three groups of experimental data show that: the farther away from the detection position,the more serious the signal attenuation.The signal strength can be significantly enhanced with the increase of the input amplitude when the underwater speaker equipment is at the same frequency.When the input amplitude is constant,the higher the frequency,the weaker the signal intensity.(3)In the case of weak characteristic signals,the characteristic signal is not obvious and easy to drown in the background when the collected data was processed directly by Fourier transform.The noise can be reduced effectively,and the characteristic signal was enhanced obviously when the experimental data were further processed by filtering and wavelet analysis.(4)The spectral analysis can only get the relative intensity distribution of each frequency component at a certain time,in order to observe the frequency variation with time in a period of time,time-frequency analysis methods such as short time Fourier transform,spectrum and Wigner-Ville distribution are applied in this paper.In order to solve the problem of the existence of cross terms in the bilinear time-frequency analysis method,the method of adding window function is used to restrain the cross terms.In order to increase the time-frequency aggregation,the data was processed by the rearranged time-frequency distribution.The system uses optical fiber type experimental device to avoid the exposure of light path in the air,greatly reducing the interference of the external environment.