Research On Ground Penetrating Radar Signal Preprocessing And Imaging Techniques

Ground penetrating radar(GPR)uses the principle of refraction and reflection of electromagnetic waves in the different electromagnetic properties on the surface of the object boundary,to reconstruct the information of underground scenes.Due to the complex GPR detection environment,echo signals not only exists in the effective signal components,but also the direct coupling wave components between transceiver antenna,the direct emission wave components,noise components and radio interferences,etc.Without GPR signal preprocessing,the efficient signal components are easy to be submerged in the powerful direct wave components,so it's difficult to effectively detect the targets.According to the GPR echo signals model,several signal preprocessing methods have been studied for the direct wave removing and noise suppression in this paper,the validity is evaluated using both simulated and real GPR data.GPR imaging is a signal processing technology,which displays the underground scenes in a more intuitive way by exchanging the informations of underground target echos into image informations.Due to the complexity of the layered medium scenarios and bistatic models,an in-depth research has been carried out in this paper in the aspects of time domain back projection imaging technique,frequency domain range migration imaging technique and compressive sensing imaging technique.Among them,according to the problem of high computational complexity to BP algorithm in GPR scenes,some enhanced BP algorithms with the calculation efficiency have been carried out.At the same time,according to the problem of high sidelobe and clutter energy,an enhanced BP algorithm based on cross-correlation information has been proposed in this paper,and combining lookup BP algorithm to improve the performance in calculation efficiency and the sidelobe and clutter suppression,the validity of the proposed algorithms is evaluated using both simulated and real GPR data.Then,in view of the limitations of RM algorithm in GPR scenes,layered RM algorithm(LRM)and bistatic LRM(BLRM)have been studied in this paper,the validity of the proposed algorithms is evaluated using both simulated and real GPR data.Finally,based on compressive sensing(CS)theory and sparse GPR imaging model,three CS imaging algorithms have been researched in this paper.Compared to the traditional algorithms from the point of focus degree,the CS imaging algorithms are better ang more valid,and are researched from the success probability of imaging position with signal-to-noise(SNR),the compressed data length and interval between two targets.