Deconvolutional Fourier-splines Based Method for Geophysical Data Conversion With Compression
Event name: Geomodel 2018
Session: Прикладные вопросы обработки сейсмических данных
Publication date: 10 September 2018
Info: Extended abstract, PDF ( 407Kb )
Price: € 20
The material of the report continues the direction of the reports on "Geomodel-2008/2009/2016/2017". As a result of the applied investigations based on the new functions class "Fourier-splines" the innovative approach to the data representation and saving and also compression excluding weak components was introduced. The solution is especially actual in geophysics where particularly seismic cubes have the amount up to tens Gygabytes (and even more). The regular equidistant data conversion is got by the mostly simple algorithm as a couple of recursive filters with alternating coefficients. In general case (prognosis maps synthesis as an example) the solution is implemented based on additionally optimized the conjugated gradients method. The method is especially efficient via the hardware acceleration (parallel processing GLSL/NVIDIA CUDA). Consequently the universality of the proposed method for arbitrary dimensions data (1D - seismotraces, 2D - profiles, horizons and maps, 3D - seismic cubes) the further perspective development of the general digital standard is carrying out (incuding digital HDTV and DVD).