The strong noise produced by the leakage of electricity from marine seismic streamers is often received with seismic signals during marine seismic exploration. Traditional denoising methods show unsatisfactory effects when eliminating strong noise of this kind. Assuming that the strong noise signals have the same statistical properties, a blind source separation （BSS） algorithm is proposed in this paper that results in a new denoising algorithm based on the constrained multi-user kurtosis （MUK） optimization criterion. This method can separate strong noise that shares the same statistical properties as the seismic data records and then eliminate them. Theoretical and field data processing all show that the denoising algorithm, based on multi-user kurtosis optimization criterion, is valid for eliminating the strong noise which is produced by the leakage of electricity from the marine seismic streamer so as to preserve more effective signals and increase the signal-noise ratio. This method is feasible and widely applicable.
Spectral analysis shows a low-frequency shadow under the BSR interface. Traditional low-frequency shadow analysis is based on stacked data. In order to understand the BSR low-frequency shadow more clearly, a frequency division analysis on stereoscopic observation seismic data based on the adaptive optimal-kernel （AOK） frequency analysis method is presented. It includes ocean-bottom seismometer （OBS） data （common receiver point data including vertical and horizontal components）, vertical cable data （common receiver point data） and horizontal cable data （stacked section of different offsets）. The OBS data frequency analysis gets a conclusion that vertical component has a significant effect on the low-frequency shadow, but the horizontal component did not. The vertical cable data shows that the low frequency band of vertical cable is wider than OBS. And then the horizontal cable data frequency analysis points out that the bigger the angle of incidence is, the more obvious the low-frequency shadow will be. The low-frequency shadow feature is shown in the stereoscopic observation field and the visual effect on com- mon reception point data is better. The lateral reservoir distribution characteristics are predicted from low-frequency shadow feature analysis of the hydrate BSR based on stereoscopic observation.
Xueqin LiuHuaishan LiuLei XingYanxin YinJianhua Wang
