基于算子外推的非因果滤波技术的应用研究

Application of Noncausal Filtering Technique Based on Operator Extrapolation

f-x(频率-空间)域预测滤波技术是消除叠后地震记录中随机噪声的常用方法,然而该方法存在两点不足:一是保幅性较差,二是无法有效提高高频段信噪比,制约了地震资料分辨率的提高。对此,重点研究了非因果滤波器和算子外推两大技术,提出了一种基于算子外推的非因果滤波技术。实际资料的处理结果表明,该技术可明显降低反射信号的畸变,并能有效提高高频段的信噪比,通过后续的谱白化等高分辨率处理技术可明显改善资料的分辨率,结果令人满意。

Prediction of filtering in f-x (frequency-space) domain was conventionally used for eliminating random noise in poststack seismogram data,but there existed two shortcomings in the method,such as poor amplitude protection and without effective improving the signal-noise ratio,by which the seismic data resolution was constrained.In consideration of the shortcomings,the noncausal filtering technique and operator extrapolation method were studied,a noncausal prediction filtering technique based on operator extrapolation was presented.The results of seismic data process show that this method can effectively protect the available signal and improve high-frequency data signal-noise ratio.It has good effect on improving seismic data resolution in combination with spectral whitening.