基于二阶谱及多阶微分融合的频谱拓展方法

Seismic bandwidth broadening method based on bi-spectrum and multi-order differential fusion

受地下介质吸收衰减及环境噪声的影响,地震反射波的频率主要集中在中低频,且频带较窄,信噪比较低.反褶积方法是解决此类问题的重要手段,而谱模拟反褶积方法因克服了传统反褶积的假设条件"反射系数为白噪声"而备受推崇,但子波振幅谱无法准确获取的缺陷限制了其应用.为此,在传统的谱模拟反褶积方法基础上进行改进,提出了一种基于二阶谱及多阶微分融合的频谱拓展方法.微分算子在频率域具有单调递增的线性特征,可提高信号的高频成分并压制低频成分,即具有分频属性,且反映频率的高低与微分阶数成正比.以期望子波振幅谱为约束条件,对不同阶微分信号的振幅谱进行融合,可获得精确的宽频地震资料,避免传统谱模拟反褶积方法在求取子波振幅谱过程中存在的误差.经过对薄互层模型及实际地震记录的试算,获得了比传统方法更好的效果,证明本文方法对提高地震资料分辨率比传统谱模拟反褶积方法更加有效.

Affected by absorption and attenuation of underground media and environmental noise, the frequency band of seismic reflected wave is narrow and mainly concentrated on the middle and low frequency and SNR is relatively low. The method of deeonvolution is an important way to solve these problem. But they have several hypothetical conditions. Among many types of this method, spectral deconvolution has been praised for overcoming the defect of the traditional methods. It does not need hypothesis that reflection coefficient sequence is white noise. But an fatal limitation, which it is not easy to calculate the wavelet amplitude spectrum, limits its application. For this reason, an improved method, which called Seismic bandwidth broadening method based on bi-soectrum and muhi-order differential, is out forward on the basis of the traditional method of spectral deconvolution. Differential operators have linear features of monotone increasing and it can improve the high frequency and suppress the low one contrarily when it is used for signal processing. Actually it has attribute of frequency spectral decomposition and higher differential order represents high frequency. Now we acquire differen! signal spectra after differential processing. The next step is combining these different spectra under the constraint of expected wavelet amplitude spectrum. In this process, it is necessary to establish a mapping relationship between predicted seismic data amplitude spectrum and predicted wavelet amplitude spectrum, and the other one between differential amplitude spectra and the envelope of differential amplitude spectra. Then a fusion coefficient is obtained and it is used to the process of fusion. Finally we acquire accurate seismic data with broadband. This method can avoid the error in the process of obtaining the amplitude spectrum of the wavelet by the traditional method. It achieves better results than the traditional method through the treatment of thin interbedded model, real seismic data and improves the resolution of seismic data obviously. The amplitude spectra of the seismic data after differential spectra fusion follow the trend and behavior as the input seismic data, but with the high frequencies boosted. And the tests on real seismic data demonstrate it is more effective than the traditional methods.