Research Of GPR Signal Based On Wavelet Transform And Compressed Sensing

Ground penetrating radar(GPR)is widely applied in many fields,but the collected echo signals are seriously disturbed by noise,which is to the disadvantage of the extraction for radar image features.Meanwhile,the thickness of the pavement layers of GPR echo signal is estimated by using delay estimation.However,the weak echo signals are easily masked by the strong echo signals when the layers are greatly thin.According to the two problems,this paper mainly uses wavelet transform and reconstruction algorithm of compressed sensing(CS)to analysis and research.Firstly,this paper introduces the basic theory and denoising principle of wavelet transform.And it discusses the echo signals' denoising parameters,respectively research the wavelet basis function,choice of decomposition layers,threshold selection and threshold function selection,and carries out a mass of simulation experiment.Then,this paper proposes a new threshold function due to the disadvantage of the existing threshold functions,and the simulation experiments are used to verify the denoising effect.In this paper,it uses the traditional matched filtering algorithm to estimate time-delay,and then uses the stagewise orthogonal matching pursuit(StOMP)reconstruction algorithm of the CS for time-delay estimation on the issue of the weak echo signals are easily masked,solving the problem of sidelobe in the matched filtering method.After that,the parabola interpolation,polarization interpolation respectively combine with the StOMP algorithm for the time-delay estimation in order to improve the estimation accuracy further.The simulation results show that the combination of polarization interpolation and StOMP algorithm more effectively improves the estimation accuracy.Finally,in this paper,it proposes an improved algorithm based on the matching filter and St OMP algorithm: Gridless—StOMP algorithm,which combines with the polarization interpolation technique as well to improve the estimation accuracy of echo time-delay.