Full waveform inversion of common-offset ground-penetrating radar based on a special source wavelet and multiple integral wave-field transform

Ground-penetrating radar (GPR) can obtain the geological information by using the electrical difference between the abnormal body and the background field. Full waveform inversion (FWI) can process radar data and obtain a more accurate distribution of relative permittivity. However, when the electrical parameters between the anomalous body and the background field are quite different, the wavefield information is seriously disturbed. In that case, the FWI method can not obtain accurate physical parameters, which is worse for common-offset GPR because of the lack of illumination angles and of wavenumber coverage. To solve this problem, a novel GPR FWI method is proposed to improve the stability of inversion using a special source wavelet and multiple integral wave-field transform. The manuscript is composed as follows: first, a special wavelet is used to improve the gradient in FWI. Then the integral wave-field transform can provide low-frequency information of data. This process can make full use of the low-frequency components hidden in the original data, obtain the large-scale relative permittivity distribution structure, and reduce the instability of the full waveform inversion. Finally, three simulation examples and a practical example verify the effectiveness of the method. Especially when there is a large difference between the dielectric parameters of the abnormal body and the background field (more than ten times or even dozens of times), the proposed FWI algorithm can still obtain stable results.