Study On Adaptive And Interactive Methods For Suppressing Internal Multiples Based On Morphological Filtering

Internal multiple is a difficult problem in seismic data processing,which will not only affect the stack and migration results,but also cause errors in the results of attribute inversion.Therefore,inter nal multiples suppression is a very critical processing step.According to the difference of the actual data,this paper proposes two kinds of internal multiple suppression methods.The first method is completely adaptive,while the other method requires human-computer interaction to remove.Both methods are based on morphological filtering,because morphological filtering is based on noise waveform to remove and has the characteristics of amplitude preservation,less dependence,and "what you see is what you get",which not only does not affect the original signal,but also appears more "lightweight" compared with the traditional matching subtraction and filtering method.Adaptive internal multiple removal method is also a method of first prediction and then removal.It is mainly composed of adaptive prediction and adaptive removal,aiming at the data with no obvious difference in velocity between internal multiple and primary wave.In the first part,the adaptive prediction is based on the virtual event method,which uses the radon transform-based adaptive stratification method to stratify the data,and then uses the virtual event method to make the prediction.The second part is adaptive removal based on morphological filtering,which calculates the time interval curve of internal multiple waves through radon transform,and then uses morphological filtering to remove the internal multiple waves in the origi nal data according to the time interval curve.This method is completely data-driven,and overcomes the defect of manual participation in data delamination,and improves the efficiency of use and calculation.The interactive method is based on multiple velocity correction,which is mainly aimed at the data that the internal multiples and the effective waves are clearly distinguished in the velocity spectrum.Firstly,it picks up the velocity advance of multiple waves,and then the event of multiple waves is leveled by NMO.Then,it uses morphological filtering to remove multiple waves in the horizontal direction.Finally,the data after inverse NMO is the data of multiple waves removal.