利用波原子分解系数自适应Wiener滤波方法压制地震数据随机噪声

Utilizing the method of wave atom decomposition coefficient with adaptive Wiener filtering to suppress seismic random noise

随机噪声是地震勘探数据中常见的背景噪声,会对有效信号产生负面影响,有必要研究更适用的方法去除随机噪声。为此,提出基于波原子分解系数自适应Wiener滤波,并对滤波后的系数进行波原子重构的方法,压制随机噪声以突出有效信号。针对人工合成地震记录,应用SVD阈值、二维多尺度小波分解阈值和波原子分解阈值三种滤波方法与所提方法进行对比测试及分析,可见所提方法在压制随机噪声的同时能充分保留有效信号,使反射同相轴更清晰且连续性更好。利用该方法还对实际地震记录进行了滤波处理,并与用上述三种阈值和时频峰值等滤波方法处理后的记录作对比,进一步表明所提方法在压制随机噪声与保持有效信号方面的性能更优,具有较强的实用性。

Random noise is a common background noise in seismic data,which can produce nonnegligible influence on valid signals,and it is necessary to develop better methods to eliminate the random noise.For this purpose,this paper proposed a method that applied adaptive Wiener filtering to the coefficients obtained by wave atom decomposition and then processed the filtered coefficients through wave atom reconstruction to suppress random noise and highlight valid signals.Based on synthetic seismic records,contrast experiments and result analysis were performed among the filtering methods,including SVD threshold,2D multi-scale wavelet decomposition threshold,wave atom decomposition threshold,and the proposed me-thod.It demonstrates that the proposed method can suppress random noise while preserving valid signals,obtaining clearer reflection events with better continuity.Finally,this method was adop-ted to filter a field seismic record,which was then compared with the filtered records by the above three thresholds and time-frequency peak filtering.It verifies that the proposed method is superior in terms of random noise suppression and valid signal preservation,with strong practicality.