双谱信号重构技术在重力数据去噪中的应用  

Application of bispectrum signal reconstruction technology for gravity data de-noising.

作　　者：俞建宝[1,2] 陈高[3] 李正文[2] 

机构地区：[1]中国石油化工股份有限公司石油勘探开发研究院南京石油物探研究所,江苏南京210014 [2]成都理工大学,四川成都610059 [3]中国石油化工股份有限公司勘探南方分公司,四川成都610041

出　　处：《石油物探》2009年第3期307-313,18,共7页Geophysical Prospecting For Petroleum

基　　金：国家高技术研究发展计划(863)项目(2008AA06Z202)资助

摘　　要：一个实信号可以用一个谐波信号序列来拟合,计算实信号的双谱,通过Fourier变换推算出谐波信号的系数和相位角与实信号双谱振幅谱和相位谱之间的关系,可以实现信号重构。谐波基频的准确估算是信号重构的关键,经过试验分析,取双谱振幅谱的第一个峰值对应的频率作为估算基频。构成实信号的谐波项数是未知的,通过设立系数因子控制谐波项数,可以满足不同频率(或波数)段信号的拟合,实现噪声压制。数值模拟和实际应用结果表明,双谱信号重构技术不仅能有效压制随机噪声,还可以进行异常分离。A real signal can be fitted by a harmonic signal series to calculate the amplitude spectrum and phase spectrum.The relationship between the coefficient,phase angle of the harmonic signal and the amplitude spectrum,phase spectrum of the bispectrum is calculated by Fourier transform.Then,the signal reconstruction based on bispectrum is accomplished by this relationship.The estimation of fundamental frequency in harmonic wave is the key for this method.By test analysis,we take the corresponding frequency to the first peak value of amplitude spectrum as estimating fundamental frequency. The number of harmonic wave for constructing actual signal is unknown. By setting coefficient factor to control the number of harmonic wave,the fitting of signal with different frequency(different wave number) can be satisfied,and the noise can be suppressed. Numerical simulation and actual application result indicate that bispectrum signal reconstruction can effectively suppress random noise,as well as do anomaly separation.

关 键 词：双谱分析 信号重构 重力异常 去噪 异常分离 归一化总梯度 

分 类 号：P631.4[天文地球—地质矿产勘探]