改进的1范数匹配滤波法及在南海深水盆地的应用

地震勘探中多次波的存在会影响速度模型的建立、地震成像效果、反演和解释过程中的介质属性提取等,因此多次波的压制成为至关重要的研究课题.本文采用改进的1范数匹配滤波法,基于1范数对大值条件的开放性,利用预测出的多次波模型Hilbert变换道及求导道,从理论上克服2范数的大值条件,并改善常规1范数下的正交性条件,在一次波和多次波叠合之处及能量差异大的情况下有效压制多次波.模型及南海深水盆地实际海洋地震数据的处理结果显示了改进的1范数匹配方法在压制多次波方面相对于常规2范数、1范数的优越性.

An improved matching method based on L1-norm for multiple elimination and application in deepwater basin of South China Sea

Research on multiple elimination has been a vital subject,because its existence can influence the building of velocity model,the effect of imaging,the extraction of attributes during inversion and interpretation,and so on.This paper puts forward an improved matching method for multiple elimination based on L1-norm without big value constraining condition,and makes use of Hilbert transformation and derivation to get phase angle adjustment for different degrees.This improved method theoretically overcomes the big value condition of L2-norm,and ameliorates the orthogonal constraint condition under general L1-norm through multi-original channels and multi-transformed multiples.With this method,multiple waves are eliminated and primary reflections are retained under the intersection and great disparity between primary reflections and multiples.The results of Pluto model and field data from deepwater basin of South China Sea declare the improved L1-norm matching method is superior to general L2-norm,general L1-norm matching methods.