Research On Several Problems Of PGPR Echo Data Processing

On the base of analyzing conventional seismic data processing method, Ground Penetrating Radar (GPR) processing method and specific working mode of Phased-array radar, the dissertation investigates several problems in Phased-array Ground Penetrating Radar (PGPR) echo data processing. This is originated from "Research on Signal Processing and Visualization for Phased-array Ground Penetrating Radar (PGPR) (Project Number: 2001AA132050-03)", which is a sub-project of the Nation 863 Project "Exploring Technology of High ResolvingPower Ability for Penetrating Surface Layer--Phased-array Ground PenetratingRadar" (Project Number: 2001AA132050).Firstly, the working mode of PGPR and the displaying mode of PGPR echo data are introduced and several simplified model are set up. For example, regarding the sending array as a sending point is the necessary and reasonable hypothesis for following data processing, and setting up ideal layered-medium model is the best means to validate various processing methods. The comparing of PGPR's observation mode with others' establishes solid theory foundation to deduce PGPR processing methods by using seismic processing method, Ground Penetrating Radar (GPR) processing method for reference.Secondly, the dissertation applies Short Time Fourier Transform (STFT) and Wigner Transform to PGPR stratified data (provided by Jianghan Petroleum Institute), studies the influence of various methods, parameters and receiving units on time-frequency diagram and presents parameters suitable for delamination. Then the true PGPR echo data (collected by XBX-001 PGPR at Sanxia dam in Dec, 2003) are analyzed by time-frequency methods and STFT algorithm is improved according to the characteristic of PGPR echo signal, and the suitable wavelet is chosen to analyze echo data. The experimental result shows that the application of time-frequency analysis, especially improved STFT and Wavelet Transform to PGPR signal processing is effective.In addition, the removal of direct wave in shallow layer underground acoustic signal and PGPR echo is studied. But due to the differences between acoustic wave and electromagnetic wave as well as different collecting modes of two models, themethod that can remove direct wave in shallow layer underground acoustic signal making use of the discrepancy of apparent velocity is not suitable for PGPR echo data. According to characteristic of direct wave in PGPR echo, 2D filtering, matrix filtering and horizontal high pass filtering are tried to remove direct wave in PGPR echo. We also use matrix filtering to process profiles and introduce adaptive matrix filtering into it to make the object in profiles more clear. Moreover, a kind of hyperbola suppression filtering is designed and experimental result shows its efficiency on suppressing diffracted waves in PGPR profiles.Finally, the dissertation does deep research on the migration of PGPR echo data. The choosing of migration velocity is analyzed in detail and the concepts of migration stacking and stacking migration are distinguished from each other. According to the working mode of PGPR, conventional diffraction scan migration in seismic processing is unproved. Basic principles of finite-difference migration and diffraction scan migration are presented and the processing results of the practical GPR data with these two migration methods are compared. Migration is applied to GSSI data (collected by SIR-10 GPR) and true PGPR echo data and it is illustrated that migration can refocus the reflecting points in radar record to their true spatial location and provide the foundation for interpretation, thus improving precision of interpretation of echo profiles.