Research On Acoustic Signal Analysis Method Of Target Field Based On Wavelet Theory

Wavelet theory has property of time frequency localization.And it is suitable for processing transient and non-stationary signals.Based on the wavelet theory,the acoustic signal acquired in acoustic target system is processed and analyzed.The design of the lifting wavelet algorithm is completed in two aspects:acoustic target signal denoising and characteristic wave recognition.The effectiveness of the lifting wavelet algorithm is verified in Matlab.At the same time,lifting wavelet algorithm is redesigned by use of verilog language facing to FPGA in order to improve the real-time quality and measurement accuracy of acoustic target system.The main work of this paper includes the followings:(1)According to general characteristics of acoustic signal,current analysis theories is introduced and compared in order to choose the proper analysis theory.(2)According to characteristics of acoustic signal,the lifting wavelet algorithm is designed,which mainly includes acoustic signal denoising and characteristic wave recognition.The selection of the wavelet function,the setting of the denoising threshold and the setting of the corresponding decomposition number are completed in denoising module.And spectrum entropy algorithm based on wavelet is used to extract the characteristic wave and identified the characteristic value of the target acoustic signal.(3)Based on wavelet theory,according to the selected wavelet function,the mathematical structure of the chosen lifting wavelet function is deduced and the lifting step is introduced.Through the process of mathematical analysis,the basic operation of the lifting wavelet algorithm is analyzed.(4)The various modules in lifting wavelet algorithm are realized in the FPGA hardware language.It mainly includes the design of basic modules,wavelet fifth-order positive transform,inverse transform,denoising module,characteristic wave recognition module and acoustic target coordinate calculation module.At the same time the corresponding simulations to each module are accomplished in this paper.The acoustic signal acquired in target system is processed by using lifting wavelet algorithm proposed.And signal arrival time value is taken into the acoustic target coordinate calculation model.Experimental results show that the lifting wavelet algorithm based on FPGA can improve the real-time performance of processing algorithm for acoustic the acoustic target signal.At the same time,compared with the traditional wavelet algorithm,the maximum deviation of X and Y coordinates decreased by 6.0mm and 5.4 mm respectively,the standard deviation of X and Y decreased by 3.79 and 2.02,and the percentage of X deviation and Y deviation within ±6mm improved by 11%and 13%respectively.The above indexes show that the algorithm effectively improves the measurement accuracy of the acoustic target.