Refining the amplitude of inverse scattering series internal multiple predictions

Abstract

Internal multiples are a form of noise in seismic data that degrades seismic image processing, leading to incorrect interpretations of the data. One method with minimal assumptions to attenuate internal multiples uses the inverse scattering series. Two key improvements with the method are developed. First, a new tool is established referred to as the downward generator space. This space used in combination with higher order terms and 2D adaptive subtraction improves the accuracy of the predicted internal multiples. Second, a non-stationary search limiting parameter is utilized in the time-offset domain allowing for increased flexibility in algorithm implementation. These new tools are applied to both synthetic and real data examples displaying improvement in the prediction and attenuation of internal multiples.