基于反射系数谱拟合的反褶积白噪化校正

Deconvolution whitening effect correction based on reflection coefficient spectral fitting

目前用于生产的主流反褶积类型为统计性反褶积,这类反褶积是基于反射系数为白噪声假设而建立的。该假设前提与实际反射系数不符,统计性反褶积输出存在误差。围绕该问题,从地震记录褶积模型出发,应用反射系数非白噪模型,在频率域给出统计性反褶积因子计算公式,指出统计性反褶积输出趋向白噪化并偏离真实反射系数。在此基础上,推导出统计性反褶积输出校正公式及基于反射系数谱拟合的校正实现方法。实际应用表明该校正方法是有效的。

At present,the main type of deconvolution used in production is statistical deconvolution,which is established based on the assumption that the reflection coefficient is white noise.Since the assumption does not match the actual reflection coefficients,the output of statistical deconvolution has errors.To focus on this problem,starting from the seismic record convolution model,the reflection coefficient non-white noise model is applied,and the statistical deconvolution factor calculation formula is given in the frequency domain.It is pointed out that the statistical deconvolution output tends to become white noise and deviate from the true reflection coefficient.On this basis,it is derived a statistical deconvolution output correction formula and a correction realization method based on the reflection coefficient spectrum fitting.The actual application shows that the correction method is effective.