Seismic Signal Separation Based On BP And Primary Application

In geological exploration or oil exploration, seismic exploration is an important geophysical methods, often in the soil for blasting to stimulate seismic waves, the seismic wave in the ground different geological spread, its waveform, amplitude strength and path with the Through the flexibility of different media change, using the received seismic waves to estimate the reflectance function, the reflection function and the impact of response to the different strata, inferred subsurface structure and rock properties, can achieve the purpose of exploration.Recorded by the seismic reflection coefficient is often called to solve the deconvolution operation, that is isolated from the seismic records the reflection coefficient. Deconvolution earthquake areas with sparse deconvolution of atoms have similar theoretical model. This paper discusses the seismic deconvolution and atomic contact between deconvolution and introduced the sparse areas of common methods MP and BP. Second, the sparse decomposition of the new method of signal processing, design of sparse decomposition based on the seismic signal separation algorithm, using ricker wavelet as atom model, synthetic seismic signals, use of BP and the MP on the noise-free synthetic seismic signals and seismic signals under various noise separation, BP confirmed by experiment in the earthquake in the superiority of Signal Separation. Two-dimensional seismic records for real, on every separately processed. The experiment were selected several different seismic profiles BP separated discuss the different phases of the seismic wavelet as the atomic model of the separation, through experimental results show that selected for the experiment section, zero-phase wavelets are to obtain the best separation. Finally, the paper also on the same layer of the reflection coefficient of the extraction were studied. Designed with the layer of reflection coefficient extraction algorithm research and implementation of two-dimensional artificial seismic signal in a row, with cases of non-continuous layer of reflection coefficient extraction, and the real seismic signal reflection coefficient on the same level achieved over the extraction good results.