Adaptive multiple subtraction method and device

Abstract

The invention provides an adaptive multiple subtraction method and device. The adaptive multiple subtraction method comprises the steps that predicted multiples in two-dimensional data windows and original seismic data in the two-dimensional data windows are utilized to construct feature vectors and label values correspondingly, and training data are generated according to the feature vectors andthe label values; the training data serve as input, and a support vector regression function is estimated through a support vector regression method; the feature vectors are input to the support vector regression function, and estimated multiples in the two-dimensional data windows are obtained; the estimated multiples are subtracted from the original seismic data, and primary estimation results in the two-dimensional data windows are obtained; and estimated primaries of the original seismic data are obtained according to the primary estimation results, in all the two-dimensional data windows,of the original seismic data. By attributing adaptive multiple subtraction to the support vector regression problem, protection of the primaries and separation of the multiples can be better balanced.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a multiple wave adaptive subtraction method and device. Background technique [0002] Seismic migration algorithms widely used in industry today are mainly based on the assumption that the energy of seismic waves is reflected only once underground. However, multiple reflection waves, referred to as multiple waves, are formed when seismic waves propagate back and forth between strongly reflecting interfaces. In the seismic migration algorithm based on primary reflections, the existence of multiple waves will reduce the accuracy of seismic migration, which will cause difficulties in seismic interpretation and inversion. Suppressing multiple waves reasonably and effectively is an important method to deal with multiple waves. Multiple wave suppression methods are divided into filtering method and prediction subtraction method. The filtering method uses the separability of ...

AI Technical Summary

Problems solved by technology

[0004] In view of the above problems, the present invention provides a multiple wave adaptive subtract

Images