一种智能化断层保护的随机噪声压制方法

A random noise attenuation method with automatic fault recognition

随机噪声是影响地震资料信噪比的主要噪声之一,在地震资料处理中,随机噪声的压制效果直接影响地震勘探成果的可靠性。分析认为,目前最常用的f-x域随机噪声压制方法会损失部分有效信号,特别是存在断层时,f-x域随机噪声压制方法的应用会使断层成像模糊。通常只能采取人工解释断层位置后,再进行非断层区噪声压制以保证断层信息不受损失,工作效率低,效果受人工解释的精度影响较大。为此,提出了一种智能化断层保护的随机噪声压制方法,该方法基于断层与连续介质的基本差异,设计双向预测算子和断层位置自动识别方法,然后在压制随机噪声的同时自动进行断层的加权保护,从而大大减少了传统f-x域随机噪声压制方法的断层模糊效应,并且无需人工解释断层位置,实现了智能化的断层保护机制和随机噪声压制效果。模型数据和实际地震资料测试结果表明:所提方法能够在压制随机噪声的同时自动保护有效信号,使断层清晰成像,从而提高了成果资料的可靠性。

In seismic data processing,the suppression of random noise is a necessary step for reliable seismic exploration.Conventional random noise suppression methods in the f-x domain cause the loss of some effective signals.In case faults are present,this may result in blurry fault imaging.A common workaround for this issue is to infer the position of faults manually prior to performing the noise suppression.However,this is a time consuming strategy,which is heavily affected by expert judgement.In this study,an integrated method for random noise suppression and fault recognition was proposed.The method exploits basic differences that make faults stand out from the continuous medium,based on which the two-way prediction operator and the automatic fault recognition were designed.The latter was carried out automatically while suppressing the random noise,thereby eliminating the blurry fault imaging typical of conventional random noise suppression methods in the f-x domain.The proposed method is fully automated,thus making manual fault interpretation unnecessary.Applications to both synthetic and actual seismic data showed that the proposed method can successfully preserve effective signals while suppressing random noise,thus achieving clear fault imaging and improving the overall reliability of the results.