摘要
本文在Donoho的小波域阈值去噪处理方法基础上,提出一种基于二次小波变换的地震信号去噪方法,该法是对地震信号进行多尺度小波变换,然后对主要由噪声控制的尺度1上的小波系数再进行第二次多尺度小波变换,并将二次小波变换后的尺度1上的小波系数置零,将其余尺度上的小波系数重构,最后对处理后地震剖面进行小波域阈值去噪处理。通过理论模型和实际资料的处理结果表明,本文所述方法能有效地改善地震剖面处理效果,提高地震资料的信噪比
On the basis of Donoho's wavelet threshold noise-eliminated processing method,the paper presented a noise-eliminated method for seismic signal based on second wavelet transform.The multi-scale wavelet transform is carried out for seismic signal in the method,then second multi-scale wavelet transform is carried out again for wavelet coefficients in scale 1 mainly controlled by noises,zero is set for wavelet coefficient in scale 1 after second wavelet transform and reconstruction of wavelet coefficients in other scales is carried out,finally,the wavelet threshold noise-eliminated processing is carried out for seismic section after above-mentioned processing.The results of theoretical model and practical data processing showed that the method presented by the paper can effectively improve processing quality of seismic sections and S/N ratio of seismic data.
引文
[1] DonohoD.De- noising by soft- thresholding.IEEE Tr-ans onIT,1995,3:613~627
[2] MallatS G.Characterization of signals from m ulti-scales edges.NYU,Com puterScienceTechReprot,1991
[3] MallatS G.A theory form m ultiresolution signal de-com position:the wavelet representation.IEEE TransPAMI,1989,11:674~693
[4] MallatS G et al.Singularity detection and processingwith wavelets.IEEE Trans onInformationTheory,1992,38(2):617~643