Study On Key Techniques For Seismic Data Special Processing And Interpretation Platform Basing On QT-GPU

With the rapid development of computer technology, the integration technique ofintegrated cross-platform, multi-lingual programming, parallel processing has becomethe focus for seismic processing and interpretation study. Linux system has multi-user,multi-task, multi-thread, high reliability, high stability, low-cost and open-sourcefeatures, which is the preferred operating system for seismic processing andinterpretation software development. In this thesis, the integrated parallel specialprocessing and interpretation system of seismic data has been carried out on the basisof thorough investigation and analysis, summing up and absorbing the previouselaborate on seismic data processing and interpretation methods, using lastestcomputer technology, such as software integration, Linux open source software andGPU parallel computing techniques.Based on in-depth study, a Linux-basedQT-CUDA parallel development architecture and an interactive, WYSIWYG(Whatyou see is what you get),integrated parallel special processing and interpretationsystem platform has bee proposed and implemented. The integrated system platformintegrates the latest computer software and seismic data processing and interpretationtechniques, which is not only the production platform, but also a good platform forfurther seismic data processing and interpretation research. The integrated systemplatform is applied to Iran kashan3D real seismic data by initial processing andinterpretation, good results have been achieved.There are six chapters in the thesis.Chapter one is about the review of theimportance,history,and existing problems of the integrated sesmic data processing andinterpretation system. Chapter two is about research on visualization of seismic data,by using2D bilinear interpolation algorithm, jumping points quadrilateral fastdrawing algorithm and color image GPU parallel rendering algorithm,an interactivevisualization of2D seismic data has been carried out, Which has greatly optimizedvisual performance. Chapter three is about arbitrary data extracting methods for seismic data special processing and interpretation, such as3D SEGY arbitraryprofile data extraction algorithm,time slicing algorithm, arbitrary horizon extractionalgorithm, arbitrary data source window extracting algorithm have been carried out.Chapter four is about2D seismic data special processing and interpretation methodsand implementation, which focus on the spectrum of seismic data,zero-phase filtering,data smooth processing, wave equation post-stack migration methods, such asfrequency domain phase shift time migration and phase shift plus interpolation of thedepth migration. To facilitate the work of reservoir prediction, a variety of seismicattribute extraction methods have been carried out based on Hilbert transformalgorithm. Chapter five is about system integration of seismic data special processingand interpretation methods,which includes an integrated QT and CUDA paralleldevelopment architecture, arbitrary data sharing, arbitrary reusing between processingand interpretation modules, calling an external FORTRAN program directly andreal-time visualization.Chapter six is the conclusions and recommendations.The main innovations in the thesis are as followings1.The arbitrary data extracting methods of the integrated system platform have brokenthe source code protection in commercial software. Besides, the arbitrary data sourcewindow extracting method within the target of seismic profile is proposed andimplemented.2.The real-time visualization of seismic data processing and interpretation results isproposed and realized by programming,which is a very good WYSIWYG approach.The approach includes variable area record section of jumping points quadrilateralfast drawing algorithm and color image rendering algorithm of the GPU parallelrendering algorithms.3.Data sharing, integration, arbitrary reusing between processing and interpretationmodules are proposed and implemented by programming. Moreover, the effects ofdifferent processing and interpretation methods based on original or processed datacan be immediately compared.4.The interface method of directly calling an external FORTRAN program isproposed and implemented by transfering data format of FORTRAN.5.To achieve the GPU parallel computing, the QT-CUDA parallel developmentarchitecture is proposed and implemented, which can carry out the GPU seismic dataparallel processing and interpretation based on QT platform. In the thesis,applications of GPU technology can not only greatly improve the speed of visualization, but alsomake the seismic data parallel processing and interpretation realized, which is a solidfoundation for further study.