Research On Wavelet Threshold Denoising Theory And Its Application In GPR Detection Signal Of Tunnel Lining

Ground penetrating radar(GPR)method stands out from many detection methods because of its fast,non-destructive and efficient characteristics,and has become the main means of tunnel lining quality detection.In the actual engineering detection process,due to the influence of various factors,the echo signal received by the wide-band GPR contains a lot of noise,which seriously affects the subsequent interpretation work.Wavelet analysis is a powerful tool for signal processing,so this paper combines wavelet threshold denoising method to study this problem.Firstly,the current situation of wavelet threshold de-noising method is studied,and the wavelet threshold de-noising method is improved by theoretical analysis,simulation experiment,indoor experiment and other research means.A new method of GPR signal analysis of tunnel lining based on the improved wavelet threshold method is proposed in combination with modulus maximum method,which can effectively improve the accuracy of GPR signal analysis of tunnel lining.The main work and research results completed are as follows:(1)Through the analysis and research of wavelet threshold de-noising method,it is found that there are still some problems in wavelet threshold de-noising at present.The first one is that there is no clear basis for how to select the appropriate wavelet base;the second one is that when selecting the threshold value,generally all scales choose the same threshold value,but with the increase of signal decomposition scale,the amplitude of signal increases and the amplitude of noise decreases,it is unreasonable to use the same threshold value;the third one is that some limitations of the traditional threshold function make the denoising effect unable to meet the current accuracy requirements.In order to solve these problems,the selection of wavelet basis function,threshold selection formula and threshold function are optimized in the following research.(2)The construction of composite index provides a clear basis for selecting the most suitable wavelet base for tunnel lining signal.By analyzing the properties of wavelet and combining with the characteristics of GPR signal of tunnel lining,the wavelet bases suitable for GPR signal analysis of tunnel lining are daubechied wavelet,coilets wavelet and Symlets wavelet.By constructing three new indexes to solve the limitations of traditional indexes,and then considering the physical meaning and digital characteristics of the three indexes synthetically,a composite evaluation index is constructed to solve the problem of multi index conflict,and the optimal wavelet base is selected from the above wavelet system.Finally,the feasibility of the composite evaluation index is verified by simulation experiment.(3)A new threshold selection formula and a new threshold function are constructed.This paper studies the typical threshold and the traditional threshold function.By analyzing their limitations,the new threshold selection formula and the new threshold function are constructed.The new threshold selection formula introduces the function parameters,which makes the threshold gradually reduce with the increase of the scale,more in line with theactual situation.The new threshold function has good continuity,convergence,and relatively smooth transition near the threshold Band.Finally,the simulation experiment is carried out,and the comparison between the visual evaluation standard and the objective evaluation standard proves that the denoising effect of the new threshold selection formula and the new threshold function is better than the traditional threshold function.(4)The improved wavelet threshold denoising method is applied to GPR signal of tunnel lining,and a new method of tunnel lining signal analysis based on improved wavelet threshold method is proposed in combination with modulus maximum method.By setting up the indoor experiment to simulate the tunnel lining structure,the new analysis method is used to analyze the detection signal from qualitative analysis and quantitative analysis.The experimental results show that the new method can remove the noise better than the traditional soft and hard threshold method,measure the thickness and range of holes and steel bars more accurately than the radar software,and effectively improve the quantitative analysis accuracy of GPR detection signal.Finally,the application effect of the new analysis method in the actual signal processing is further verified by an engineering example,and the feasibility and superiority of the new analysis method are highlighted.