Ground Penetrating Radar Echo Signal Denoising Based On Wavelet Threshold And EMD Method

Daily inspection of road quality is an effective means to prevent road collapse accidents.In view of the limitations of non-destructive testing methods,ground penetrating radar(GPR)is widely used in modern road testing due to its high efficiency and non-destructive advantages.However,due to the interference of noise,it will affect the effective signal extracted by the ground penetrating radar,thereby affecting the detection result and increasing the hidden danger of road safety.Therefore,this article focuses on the research on denoising methods of ground penetrating radar echo signals.The main work and results are as follows:1.Aiming at how to improve the effect of ground penetrating radar signal in wavelet threshold denoising,this thesis makes appropriate selection and improvement on the four factors that affect the effect of wavelet threshold denoising.First of all,for the selection of wavelet base and decomposition level,this thesis selects the db5 wavelet base and 5-level decomposition level suitable for ground penetrating radar echo signal through analysis and comparative experiments.Aiming at the problems of complex local threshold calculation and inaccurate threshold estimation in commonly used wavelet threshold estimation methods,this thesis proposes a local threshold estimation method based on noise characteristics based on the Sqtwolog threshold and the performance of noise in wavelet decomposition.Improved the simplicity and accuracy of threshold calculation.Finally,through the analysis of the advantages and disadvantages of commonly used threshold functions,this thesis proposes a threshold function with parameters in the form of dual thresholds.While the threshold function has continuity and asymptotic properties,it has also been improved by increasing the adjustment parameters and the design of the dual threshold form.The effect of threshold processing.Finally,combining the above,an improved wavelet threshold denoising method is proposed in this thesis,which has a good denoising effect.2.Aiming at the problems of modal aliasing effect,inaccurate judgment of noise components and improper processing of noise components in the partial reconstruction method of empirical mode decomposition(EMD),this thesis proposes corresponding improvement methods respectively.Firstly,the ensemble empirical mode decomposition(EEMD)and the complete ensemble empirical mode decomposition(CEEMD)were used to decompose and compare the ground penetrating radar echo signal.According to the experimental results,CEEMD with better decomposition performance was finally selected to solve the problem.Modal aliasing problem.Secondly,this thesis proposes a noise component judgment method based on the correlation of adjacent echoes of ground penetrating radar.This method uses the principle of ground penetrating radar and the correlation between effective signals when processing,and improves the accuracy of the judgment result.In addition,according to the statistical characteristics of noise,this thesis proposes an improved sort-accumulate-average noise component processing method.This method overcomes the limitation of common noise component processing methods that ignore the effective signal,so as to extract more noise components.Effective signal.Finally,combining the above-mentioned improved methods,an improved CEEMD signal denoising method is proposed.Experiments show that this method can significantly reduce the influence of noise on the effective signal.3.Aiming at the limitations of a single processing method,this thesis proposes a WT-CEEMD-SVD ground penetrating radar echo signal processing method.First,the improved wavelet threshold denoising method and the improved CEEMD signal denoising method are combined,and the WT-CEEMD signal denoising method is proposed.Through comparative experiments,the effect of the WT-CEEMD signal denoising method is verified,and the rationality of combining the improved wavelet threshold denoising method with the improved CEEMD signal denoising method is also proved.In addition,based on the analysis of the characteristics of the ground penetrating radar echo signal,this article adds a singular value decomposition(SVD)method to the WT-CEEMD signal denoising method to achieve effective suppression of clutter.Finally,through the comparison experiment with the commonly used ground penetrating radar processing method,the actual measurement data is verified to verify the effectiveness of the WT-CEEMD-SVD ground penetrating radar echo signal processing method.