Adaptive Multiple Attenuation Based On Curvelet Transform

In seismic data processing,multiples have negative influences on velocity model building,seismic imaging,structure interpretation and attribute extraction.Therefore,it is significant to eliminate all kinds of multiples effectively and reserve the primary possibly in seismic explorationThe conventional multiple adaptive subtraction method is based on least squares principle,and the algorithm cannot attenuate multiple that overlaps the primary.Therefore,we need to use other methods to improve the adaptive subtraction method.The paper focuses on the principle of Curvelet transform,and selects the appropriate matching parameters to match primary and multiple with principle of least squares in Curvelet domain to attenuate multiple.The paper first introduces the principle and characteristics of Curvelet transform,and demonstrates the method can separate multiple from primary when they are not vertical.The synthetic data examples show that this method can separate multiple from original data.At the same time the length of the filter is tested to achieve a good result.When applying the method to the read data,it doesn't show a result as good as synthetic data.The reasons and suggestions have also been given for the real data case.The method is used to process several synthetic data and field data,and the paper puts forward some suggestions on improving the method.