基于可调因子Gabor小波变换的地震高分辨处理方法

High-resolution seismic processing method based on tunable factor Gabor wavelet transform

时频分析是利用地震资料识别薄层的重要方法之一,常规的时频分析方法受到固定时窗、窗函数等因素影响。为此,采用一种基于可调因子Gabor小波(Tunable Factor Gabor Wavelet,TFGW)的连续小波变换(Continuous Wavelet Transform,CWT)(简称TFGW-CWT)的时频分析方法对地震资料进行处理。该方法采用具有可调因子的Gabor小波进行变换,然后使用最小绝对值投影方法组合每个可调因子的连续小波系数,降低邻近频率的交叉干扰,提高局部时频分辨率。进一步采用非负矩阵分解(Non-negative Matrix Factorization,NMF)对时频数据体降维,得到一个低秩的特征数据结构关系体,减少了高维空间的数据冗余,凸显了高频信息,达到提高地震资料分辨率的目的。模拟记录和实际资料处理结果证实了该方法的有效性,可为薄层检测提供一种新的技术手段。

Time-frequency analysis is one of the important methods for thin layer identification with seismic data.Conventional time-frequency analysis methods are affected by fixed time window,window function and other factors.To this end,continuous wavelet transform(CWT),based on tunable factor Gabor wavelet(TFGW),TFGW-CWT for short,is used to process seismic data.In the TFGW-CWT method,TFGW transform is performed,and then the minimum absolute value projection method is employed to combine the continuous wavelet coefficients of each tunable factor to reduce the cross interference of adjacent frequencies and improve the local time-frequency resolution.Furthermore,non-negative matrix factorization(NMF)is conducted to reduce the dimensionality of the time-frequency data body to obtain a low-rank characteristic data structure relation volume.This reduces the data redundancy of high-dimensional space,highlights the high-frequency information,and achieves the purpose of improving the resolution of seismic data.The results of simulation record and actual data processing confirm the validity of the proposed method,which can provide a new technique for thin layer detection.