Research On The Method Of Improving Ground Penetrating Radar Signal Resolution

As an efficient shallow geophysical detection technology, the development of theground penetrating radar has experienced a century from the early20th century to now. Withthe further development and application of ground penetrating radar, the resolution isdemanded to improve itself higher and higher. However, the working principle of groundpenetrating radar is that the target body is detected through the ground penetrating radartransmitting and receiving high frequency electromagnetic wave to the underground. Themain work object of the ground penetrating radar is loss medium whereas the sky penetratingradar electromagnetic wave is transmitted in idea lossless medium. When the groundpenetrating radar wave transmits in the loss medium, the attenuation and dispersion can occur.It makes the data image received become fuzzy distortion and can affect the detection depthand resolution of ground penetrating radar. This makes the advantage of the high resolutionof ground penetrating radar be of restrictions and restricts the development of high resolutionground penetrating radar technology. In view of the influence of attenuation and dispersionon the radar wave, it is very meaningful to improve the resolution of ground penetrating radar.High resolution ground penetrating radar technology research involves basic theoreticalresearch, methods research and equipment three main areas. In this paper, aimed at improvingground penetrating radar resolution technology, methods research is the focus. More precisely,the study of signal processing algorithms to improve the ground penetrating radar resolutiontechnology is the focus in this paper. For improving the ground penetrating radar resolution,the main study contents involve the estimation of ground penetrating radar quality factor Q,the inverse Q filtering and deconvolution three parts. This three parts research contents areintroduced as follows:Aiming at the problem of estimating ground penetrating radar quality factor Q, thealgorithm to obtain Q value using amplitude spectral ratio has been improved. Initially, asingle spectrum is obtained by using Fourier transform. But now, time frequency spectrum isobtained by using time-frequency analysis. The Q value is estimated by using the ratio oftime frequency spectrum.The algorithm of estimating Q value by using frequency shift method is furtherdeveloped in this paper. this algorithm is validated by taking advantage of layered constant Q-model.The wavelet is confirmed by modulation frequency, Energy attenuation factor of wavelet,the amplitude and phase of wavelet in seismic theory. Drawing lessons from this theory inthis paper, the Q value in each layer is set as a constant in horizontally layered medium (Qvalue is independent of frequency). And firstly the frequency-domain equation of the singlepass propagation theory of plane wave is solved. Through definition envelope peakinstantaneous frequency and discussing algorithm, it shows that Q value is confirmed by thechanges of envelope peak instantaneous frequency, travel time and the wavelet energyattenuation factor. So the method to estimate Q value using envelope peak instantaneousfrequency of ground penetrating radar VRP data direct wave is obtained. At last, theeffectiveness and accuracy of the algorithm is verified using synthetic VRP data andsynthesis experimental data applying FDTD theory and practical data.Drawing lessons from inverse Q filtering algorithm in the seismic theory, the theory ofwavefield downward continuation is combined with generalized S transform and the jointtheory is applied to the data processing of inverse Q filtering of ground penetrating radarsignal horizontally layered medium. The underground Q model is assumed to layeredconstant Q structure and for each constant Q layer, the earth's surface wavefield records isdirectly prolonged the top of current layer; the wave domain prolonged is done generalized Stransform in time frequency domain in constant Q layer and the signal is done inverse Qfiltering; the filtering result is done generalized S inverse transform and the time domainsignal after being inverse Q filtering is obtained. Finally, the complex layered groundpenetrating radar attenuation records based on the theory of wavefield downwardcontinuation and the two layer ground penetrating radar attenuation records obtained fromFDTD theory are done amplitude and phase compensation. It verifies the effectiveness andaccuracy of the algorithm in ground penetrating radar inverse Q filtering.Inverse Q filtering method in ground penetrating radar absorption dispersion mediumand the inverse scattering theory of medium are combined in this paper. The problem ofinverse Q filtering of layered medium in ground penetrating radar is transformed into theproblem of inverse scattering of medium according to scattering domain theory and greenfunction knowledge. The distortion of amplitude and phase generated in communicationprocess of electromagnetic waves in the underground is compensated by solving inversescattering problem.The traditional ground penetrating radar convolution model is discussed in this paper. Itproposes to use linear time-varying systems to describe the radar return signal. Namely, theradar return signal is described to linear integral form between wavelet and formation. For validating the correctness of this model, the model is used in the process of extracting groundpenetrating radar wavelet in this paper. In this process, the underground medium used aretraditional layered medium and random medium possessing local small scale inhomogeneouscharacter. The final result shows that this kind of linear time-varying systems can be better todescribe the ground penetrating radar return signal than the convolution form of thetraditional linear time invariant system.Fractional Fourier transform theory is applied to ground penetrating radar waveletextraction algorithm in the process of ground penetrating radar wavelet extraction in thispaper. It assumes that ground penetrating radar signal meets the prior knowledge of fractionalFourier transform optimal filtering in the process of wavelet extraction. The problem ofwavelet extraction can be transformed the problem of optimal filtering in fractional Fouriertransform domain since fractional Fourier transform is a unitary transform and has normconserving. The optimal solution is found in fractional Fourier transform domain and finallythe time domain signal of wavelet estimation is obtained after fractional Fourier transforminverse transform. In addition, when the single optimal filter is not ideal and the signal isaffected by noise, single optimal filter can not meet the needs. Thus a concept of multilevelfilter in the fractional Fourier transform domain is proposed. The shape and characteristics ofwavelet have no special requirements in the wavelet extraction algorithm in this paper. So,the algorithm described in this article is more universal.Ground penetrating radar deconvolution technology based on formation spectrumcompensation factor and linear time varying formation system response model is proposed inthis paper. A formation spectrum correction factor in the fractional Fourier transform domainis solved using fractional Fourier transform optimal filtering algorithm to verify the originalrecords of the spectrum and improve resolution. The algorithm proposed in this paper iscompared with ground penetrating radar deconvolution algorithm built on the traditionalconvolution model and based on formation spectrum correction factor. The result verifies theproposed algorithm is more reasonable and effective than the traditional methods.