Traditionally, in surface-related multiple elimination (SRME), multiples are predicted from primaries by a spatial convolution process. However, the same relationship between primaries and multiples can be used to describe a deconvolution process, where multiples are transformed into primaries or data. This method is called a focal transform. The method that is presented in this paper is based on the focal transform and reconstructs the primaries or data in the missing near offsets by minimising the estimated primaries. The result of this method is improved by choosing a different subtraction norm and using the fact that multiples predicted along different ways should be the same.
展开▼
