Attenuation Of Multiple In Image Space

All the time, the multiple problem has always been an important issue that puzzledseismic data processing, its existence seriously affects the authenticity of the seismic imagingand disturbs the accuracy of seismic data interpretation. At present, there are mainly two waysto deal with the multiples in seismic data processing, one is considering the complex wavefield effect of multiples on primaries imaging, see it as noise to be suppressed; the other one isthat considering the fact that multiples also contains a lot of information of the undergroundreflect interface, so dealing it with imaging to assist primaries imaging, so we can obtain moreaccurate imaging result.In this paper, we discuss about the multiple attenuation, which is the most common andmature technology. Multiple attenuation is a common method to dispose multiples, and it isalso an important issue which puzzles seismic data processing.At present, multiple attenuation methods can be chiefly divided into two types: the firstone is according to the move-out of primaries and multiples or the periodicity of multiples.But this method is not suitable when the underground structure is complex, the multipleattenuation result is very poor. The second one is multiple prediction and subtractionaccording to wave theory, this method takes the propagation effect of underground wave fieldinto consideration, and it has better applicability and higher accuracy, but the calculation ofthis method is very enormous, what’s more, usually, the method needs higher requirements ondata. All of these methods are achieved in data space and have superiority and limitation.This paper introduces a new idea to carry out the multiple attenuation in image space.Firstly, do pre-stack depth migration to the seismic data to gain the angle-domaincommon-image gathers (ADCIGs). According to the different characteristic of the primariesand multiples in the ADCIGs, namely, when a correct velocity is given, the primaries andmultiples will show different angle-domain move-out in the angle gathers, the primaries areplat while the multiples are cured. So we can separate the primaries and multiples usingproper Radon transform which can be used in the angle-gathers. This kind of methods canproperly deal with complex seismic wave propagation effect under the ground, so it can bemore suitable, at the same time, it can greatly save the computation costs. In addition, through the preserved-amplitude analysis, abstract the preserved-amplitude ADCIGs, and then domatching processing to the ADCIGs in which the multiples are suppressed. After theprocessing, the ADCIGs can be better used in the following velocity analysis and AVO/AVAanalysis or other processing. In this paper, we indicate the superiority of the method used inthis paper through the model test and real data processing.