Research And Application Of Processing Technology In Low S/N Seismic Data In Talimu Basin

In the eastern plain of China, the surface geological conditions vary mildly, so the reflections in raw seismic data are relatively stable, the problems of static correction and all kinds of disturbance are not much severe and the Signal-to-Noise Ratio (SNR) is relatively high. A set of methods in seismic data processing have been established in the routine processing.However, for the seismic data from the low SNR region with complex surface, because of the complicated surface and underground geology structures and the poor transmitting/receiving conditions, there exist severe static correction problem and strong disturbance in the original records so that reflections can not be identified in raw data. The needs of geological task cannot be met with the routine seismic processing methods, which influences badly exploring and developing in this kind of areas.In the desert region of talimu, the surface is very complex. The strong lateral velocity variation of the low velocity layer (LVL) and the surface layer results in the serious problem of static correction. However, flank disturbance caused by large dune in the seismic data is very severe, which leads to the decrease of signal-noise ratio of seismic data. Flank disturbance and dispersion of high energy submerge the signals. The reflection signals are hard to be seen in some raw single-shot records.According to the features of seismic data in this region, the article emphasizes on static correction and pre-stack de-noising, which aims to eliminate static correction caused by complex dunes and some noise disturbances to increase the signal-noise ratio of seismic data.On the aspect of solving static correction problem, we adopt the matched static correction technique. Firstly, we summarize the new equations of dune curve to solve the long wavelength and medium wavelength static problem. Secondly, we adopt mutual refraction statics technique to solve the large residual statics. Finally, we adopt offset-division surface consistence residual statics method to solve the discrepancy problem of residual statics from different offset in the common reflection gathers.On the aspect of prestack de-noising, we adopt the technique of median filtering, adaptive low frequency noise suppression, and high-energy disturbance suppression to suppress the different noises respectively. Meanwhile, according to the different linear relationships and strong-weak relationships among different"fields", we apply the methods to different datasets in series and achieve a better processing results.