High Resolution Imaging Algorithm Of Impulse Ground Penetrating Radar

Based on electromagnetic wave penetrating property through non-metal medium,ground penetrating radar(GPR) can implement the detection, location and shapereconstruction of targets buried under the earth's surface. It has many advantages suchas nondestructive detection, strong penetrability, high resolution performance,operational convenience, low cost and so on. Aimed at subsurface targets twodimension and three dimension shape reconstruction, high resolution GPR imagingtechniques are studied in this thesis.In chapter 2, diffraction tomography imaging algorithm based on GPR linear arrayscanning model is studied. Based on the first order Born approximation and syntheticaperture scanning model, a three dimensional diffraction tomography imaging algorithmis educed from basic electromagnetic scattering integral equation. A brief analysis aboutthis algorithm's procedure and computation burden is provided subsequently. Under thecondition of linear array scanning, dense spatial sampling of objects' scattering field canbe obtained and this makes it possible to get higher resolution imaging results. Fromintroducing transmitting aperture vector and receiving aperture vector, a threedimensional diffraction tomography imaging algorithm based on linear array scanning iseduced and it's implementation procedure is provided. Both synthetic aperture scanningdata and linear array scanning data are collected by using FDTD method and processedby the two diffraction tomography imaging algorithm.In chapter 3, range migration imaging algorithm is studied to implement targetreconstruction when GPR is used to detect shallow subsurface targets. Unlike far fielddetection, scattering wave front is no longer planar when subsurface targets locates atGPR near field. After analyzing scattering signal's integral formula, a calibration item ofnear field spherical wave spreading factor is investigated and a near field rangemigration imaging algorithm is educed by using the principal of stationary phaseformulation. The resolution performance is also analyzed subsequently. To improvesubsurface imaging results resolution, a new signal pre-processing algorithm based onchannel calibration and pulse compression is provided. Experimental data collected by aimpulse GPR system-Radar Eye is processed by using the new signal pre-processingalgorithm and near field range migration imaging algorithm.In chapter 4, a new non-uniform spatial sampling reverse timemigration(NUSS-RTM) imaging algorithm is provided. Based on continuous spreadingproperty of scattering wave in spatial-temporal domain, a narrow angle RTM imagingalgorithm is described and its difference implementation formulation is provided. But inGPR application, narrow angle approximation and uniform spatial sampling are usuallyhard to satisfied. By using Taylor serie expanding technique, its difference discrete form of spatial derivation can be described by changing the modified factor in RTMalgorithm and thus a new NUSS-RTM imaging algorithm is established. Its resolutionperformance is also analyzed subsequently.In chapter 5, a time-delay and amplitude modified back projection(TAM-BP)imaging algorithm is established to reconstruct underground objects shape for theelevated antenna scenario, which is the usual scanning model in GPR application. Inthis scenario, traditional SAR-BP imaging algorithm will bring out serious distortion inimaging result, including scattering position distortion and scattering intensity distortion.The first one is caused by electromagnetic wave refraction phenomena at air-soilinterface and the second one is caused by spherical wave propagation loss andtransmission loss. Through analyze the factors which causes imaging results distort indetail, a new TAM-BP imaging algorithm is provided subsequently. To speed up itsprocessing, a real-time TAM-BP algorithm is established based on its recursiveimplementation model. An experimental B-Scan radargram is collected by usingRadarEye system and its pre-proeessing results and real-time imaging results are allshown in the end.