【摘要】 针对常规小波相关去噪仅在尺度域或空间域进行二阶相关去噪的局限,本文提出了一种小波尺度域和空间域高阶相关去噪方法。首先对地震记录剖面进行5个尺度的静态小波分解,得到5个分量记录剖面,对这5个剖面的每道在尺度域对相邻尺度的小波系数进行高阶相关;然后对经尺度域相关的5个剖面在空间域对同一分量记录剖面相邻道的小波系数再进行高阶相关,并用此系数对相应道修正,完成所有分量剖面的空间域相关后的重构,最终得到去噪后的地震记录剖面。数值试验和实际资料处理结果表明,该方法对随机噪声的去噪效果明显优于仅在尺度域或空间域进行二阶相关的小波去噪的效果。更多还原

【Abstract】 In view of the limitation that denoise of ordinary wavelet correlation is to use second-order correlation denoise only in scale domain or space domain, the paper presented a method for denoise of high-order correlation in wavelet scale or space domains. First, the five-scale static wavelet decomposition is carried out for seismic section, getting five component sections, and for each trace of five sections the high-order correlation is carried out for wavelet coefficient of adjacent scales in scale domain; then, for five sections after correlation in scale domain the high-order correlation is carried out again for wavelet coefficient of adjacent traces in same component section in space domain, and using the coefficient to modify the corresponding trace, which finished reconstruction after all component sections being correlated in space domain, and post-denoise seismic section is finally obtained. The numeric test and processing results of practical data showed that the method has evident denoise effect for random noise and is superior to wavelet denoise method only using second-order correlation in scale or space domain.更多还原