Marine Multi-source Blended Wavefields Separation Based on an Adaptive Iterative Multilevel Median Filtering Method

Abstract:

Blended acquisition technique can be used to improve the efficiency in marine seismic exploration, but the separation of blended wavefields is a key factor for conventional processing. When the blended record is sorted from the common shot gather to other domains, the aliasing signal of secondary source will be transformed into a random distribution noise. In this paper, we produce an adaptive iteration multilevel method to achieve multi-source separation. Based on an adaptive iterative multilevel median filtering method, aliasing noise can be removed in a large window first, but at the same time it will lost the details of the wanted wavefields. Then we can gradually extract more details by reducing the median filter window size. We test three simultaneous acquisitions for marine synthetic seismic data to verity the reliability of the method, and apply this method to the real blended data which was acquired from a marine area. The result shows that a single shot record of each source is clear after the separation, and the spectral features are similar. Thus the method in this paper is proved to have a good practical value.