Research On Signal Processing Algorithm Of Ground Penetrating Radar For Shallow-buried Small Target

Ground Penetrating Radar(GPR)is a detector that uses the reflection and scattering of electromagnetic waves produced in the discontinuities of the medium's electromagnetic properties to detect invisible targets underground.Compared with other detecting methods,it has advantages of high spatial resolution,speedy detection,the ability to detect front area and even airborne detection.With the rapid development of ground penetrating radar,it's widely applied to various aspects of military and civil uses.This paper focuses on the clutter suppression of GPR data and its synthetic aperture imaging.Main work and achievements are as follows:1.The study of signal model of GPR.Based on the A-scan signal model of GPR and combined with the characteristics of small target echo,this paper has proposed the B-scan signal model of GPR.In addition,according to the B-scan signal model of GPR,an algorithm for the location and detection of small target is proposed on basis of hyperbolic symmetry.2.The study of algorithm for clutter suppression of GPR.Firstly,this paper studies three common algorithms for clutter suppression of GPR,including averaging method,SaS(Scale and Shift)method and adaptive cancellation method.And then it makes a comparison and analysis of the performances of these three algorithms through experiments.Secondly,it mainly studies the algorithm for clutter suppression of GPR on basis of Singular Value Decomposition(SVD),makes an theoretical analysis of the SVD band pass filtering algorithm by use of inner product of matrix and the characteristics of SVD,and proposed an algorithm to calculate the cut-off point of the SVD band pass filter based on image entropy.Finally,it presents a modified SVD band-pass filtering algorithm,whose core concept is to weighting each eigenmatrix in the SVD band-pass filter,so the algorithm is also called weighted SVD band-pass filtering algorithm.For the stronger residual clutter produced during the clutter suppression of B-scan data by the SVD band-pass filtering algorithm,this paper puts forward the wavelet soft-thresholding method for image denoising to make a further suppression of the residual clutter.3.The study of algorithms for synthetic aperture imaging and wave velocity estimation of GPR.First of all,this paper introduces two algorithms for synthetic aperture imaging,which are BP imaging algorithm and F-K migration imaging algorithm.Advantages and disadvantages of these two algorithms are then compared through experiment.Next,it studies the algorithm for velocity estimation of frequency-wave number migration and the algorithm for wave velocity estimation based on Hough transform,and proposes a method of extracting target echo hyperbola from the B-scan image of GPR and an algorithm for wave velocity estimation on basis of hyperbola fitting according to the characteristics of hyperbolic function.Finally,it makes a comparison between the advantages and disadvantages of these three algorithms for velocity estimation through experiment.