Attenuation-compensated Reverse-time Migration Imaging Applied To Ground Penetrating Radar

Ground penetrating radar(GPR),as an efficient shallow geophysical method,has been widely used in civil engineering and subsurface object detection.As an important aspect of GPR data processing,imaging can help understand the position and shape of the subsurface objects.At present,GPR imaging techniques include:diffraction stacking,Kirchhoff migration,F-K migration and synthetic aperture processing.In recent years,reverse time migration(RTM)method has drew the attention of scholars all over the world because of its high imaging accuracy and that it has no inclination limit.In contrast to seismic waves,electromagnetic waves suffers from attenuation due to the conductivity and dispersion of the subsurface medium.While,the power attenuation in both the forward propagation and backward extrapolation of RTM would degrade the imaging quality.In order to mitigate the influence of the material attenuation,this paper utilizes an attenuation-compensated RTM algorithm based on the finite difference time domain(FDTD)method.During the backward extrapolation of the receiver wave field simulated by the FDTD method,the conductivity is added by a negative sign,and the electromagnetic waves are mathematically amplified.In such a way,the attenuation effect is compensated for RTM.In order to save the computation time and the memory resources,it is necessary to simplify the actual GPR transmitting and receiving antennas as point sources in the RTM imaging algorithm.However,with the purpose of ensuring the accuracy of wave field calculation,the position of the point source needs to be determined at the apparent phase center of the antenna.Therefore,this paper model four types of GPR antennas,i.e.dipole antenna,horn antenna,circularly-polarized spiral antenna and Vivaldi antenna,and determines their apparent phase centers from the simulated radiation phase in the far field.The results show that the apparent phase center of the dipole antenna has the most focused locations of the apparent phase center as the operating frequency varies,among the simulated four GPR antennas.The priority of attenuation-compensated RTM,compared with the conventional RTM algorithm,is validated through numerical and laboratory experiments.In the numerical experiments,this paper design three models with different objects,i.e.a metal ball,a glass cavity,the combination of a metal and a glass cavity,respectively buried in sand.The numerical experiments deploy 14 sources and 14 receivers at depth of 0.02m,with a horizontal spacing of 0.12m.From the numerical results,it can be seen that the first model and the second model is well imaged,the imaging of the bottom of cavity in the third model is much clearer after being attenuation-compensated.In the laboratory experiments,this paper design a model which shallow subsurface object is the metal ball and the deep target body is the glass cavity which is filled with air.The objects are buried in sand.The laboratory experiments deploy 15 sources and 14 receivers at depth of 0.02m,with a horizontal spacing of 0.12m.The model is imaged by the conventional RTM algorithm and the attenuation-compensated RTM.The paper extrapolates the zero-offset data from the CSG magnetic data,then obtain the images.From the laboratory results,it can be seen that the amplitude is recovered and the reflectors are much more cleared after being attenuation-compensated.In summing up the results of numerical experiments and laboratory experiments,this paper can conclude that the attenuation-compensated RTM can well recover the weak signals,reconstruct the original wave field and improve the accuracy of the GPR imaging.