The suppression of multiple events is a crucial task in seismicdata processing, and the adaptive subtraction of the predictedmultiples is recognized as one of the main challenges for the successof the surface-related multiple elimination technique. Thetraditional least-squares matching approach can affect the primaryevents because the estimated multiples tend to adapt tothe primaries under the minimum energy condition. We investigatetwo filtering techniques for improving the multiple removalresults. In the first proposed method, we combine the advantagesof the least-squares and pattern dip-based subtraction methods.Doing so, we exploit the separation of primaries and multiples inthe dip domain, and then we apply the least-squares adaptive subtractionin each dip band before recomposing the data to obtainthe final subtraction result. As a result of the dip decomposition,the primary-multiple interferences are reduced, allowing for amore reliable least-squares filtering. In the second method, wepropose to replace the multiple subtraction step by a separationstep using independent component analysis (ICA) methods. Weemploy the ICA method after least-squares adaptive filtering. Becauseof the non-Gaussian distributions of the involved signals,primaries and multiples can be separated by computing the optimalrotation between these two signals.We apply the ICA methodin local 2D time-space windows to better compensate thespace and time variant character of the data. Two-dimensionalsynthetic and field data examples demonstrate that the multiplesubtraction results of both methods are indeed improved withrespect to the classical least-squares method.
展开▼
