Study On Ground-penetrating Radar Multiple Waves Identification And Suppression Method

The ground penetrating radar(GPR) technology is a non-invasive superficial geophysical technology using high frequency electromagnetic waves to confirm the medium distribution. When the dielectric constant of the interface existed large difference, the receiving antenna of GPR will collected the multiple waves, and the common medium produced multiple waves mainly includes cavity, metal, pipelines,water layers and low velocity layers etc. The current GPR data acquisition using profile method, and the obtained data only representing the undulations of the subsurface reflectors, but the velocity of electromagnetic wave and other information.Therefore, in recognition of multiple waves, little information available, and the features of reflection axis can be used. There is great difficulty of recognition. The current GPR profile data often reference seismic data processing method to suppress multiple waves, but the GPR data compared to seismic data has a lot of difference, its ultra high frequency, short wavelength and different ways of acquisition made it impossible using the same way to processing. The existence of multiple waves can interfere with the identification of effective information, and the multiple reflection with strong energy will cover the effective information,if the compaction process if incomplete, it also will damage the effective information, and resulting in the interpretation of trap. With the application of GPR technology is more and more widely, the multiple waves problems are more prominent. However, there are few researches on the identification and suppression methods of GPR multiple waves at present,so it is more meaningful to study on the multiple waves identification and suppression methods of GPR.In this paper,in order to study the multiple waves characteristics of periodic and non periodic in GPR profile, we designed a lot of different geological model, and by the meas of numerical simulation we got the zero offset profile, then we analyzed the multiple waves identification characteristics in different models. In order to solve the problem of the multiple waves periods in every channel is different,the adaptive prediction error deconvolution technique is proposed. This technique is improved based on the predictive deconvolution technique, according the self-correlation result to obtained the multiple waves period. Then set the prediction step and operator length to get the initial prediction filtering factor, and using the LMS algorithm to update the prediction filter factor. And the LMS algorithm has the characteristics of low computational complexity, fast convergence speed and good stability. Compared with the predictive deconvolution method, the adaptive prediction error deconvolution algorithm has the features of multiple waves prediction is accurate and the suppression result is well. And compared with the adaptive deconvolution method, the adaptive prediction error deconvolution algorithm has the features of calculation speed fast and fewer iteration. This technology was applied to synthetic data, and confirmed the periodic and non periodic internal multiple waves and the surface-related multiple waves can be effected suppressed. Then the technology was applied to the void model physical experiment data and inclined stratum physical experiment data, confirmed the periodic and non periodic surface-related multiple waves can be effected suppressed. At last, we applied this technology in the GPR data of Badan Jilin desert dune internal structure, and successfully suppress the non periodic surface-related multiple waves.