多分辨率地震信号反褶积

MULTIRESOLUTION SEISMIC SIGNAL DECONVOLUTION

基于二进小波变换提出了一种新的反褶积方法─-多分辨率地震信号反褶积．在地震信号二进小波变换域中的各尺度上分别进行其分辨率随小波尺度变化的反褶积，利用不同分辨率反褶积结果之间的相关性，以及测量噪声随尺度的衰减特性，从低分辨率反褶积结果逼近高分辨率反褶积结果．理论分析和实验表明，该方法有较高的精度，并且在较低信噪比情况下有好的效果．

A new algorithm of seismic signal deconvolution is proposed in this paper based on dyadic wavelet transform, which is called multiresolution seismic signal deconvolution. The seismic signal is decomposed with dyadic wavelet transform, then in the domain of dyadic wavelet transform of seismic the predictive deconvolution is executed with different length of predictive step according to the different scale factor of dyadic wavelet transform. According to the correlativity between the different deconvolution results and the difference between the signal-noise ratio with different scale factors in dyadic wavelet transform domain we use the lower resolution deconvolution result to modify the higher resolution deconvolution result successively. This algorithm leads to the reflectivity coefficient with high precision, especially it is effective to the seismic signal with lower signal-noise ratio.