【摘要】 提出了地震信号谱分解的两步法匹配追踪(MP)算法。该方法根据最优小波的时间延迟,采用多点同时搜索的方式,在不损失MP算法精度的同时使运算效率得到了提高。阐述了基于Ricker小波MP算法的基本原理,给出了两步法MP快速算法的实现步骤。基于理论模型,讨论了MP算法谱分解的效果,并与Gabor变换和小波变换的谱分解效果进行了对比分析;利用谱分解的峰值瞬时频率进行了薄层反演,与常规薄层反演算法的对比表明,该方法对储层上、下界面反射系数的变化不敏感。在某地区,利用两步法MP快速算法对实际地震记录进行了谱分解,并基于由谱分解得到的不同频率的切片,分析了含气储层下方的影子效应,结果表明影子效应直接指示了烃类的存在。更多还原

【Abstract】 A two-step matching pursuit(MP)algorithm of seismic signal spectral decomposition was presented,which is based on the timedelay of the optimal wavelet,by multi-points simultaneous searching -style,to improve the operation efficiency without losing algorithm accuracy.Then the principle of MP algorithm based on Ricker wavelet was elaborated,and the processing workflow of two-step MP algorithm was given.Based on theoretical model,the effects of spectral decomposition by MP algorithm was discussed,and compared to the spectral decomposition results of Gabor transform and wavelet transform.The peak instantaneous frequency of spectral decomposition was utilized for thin-bed inversion,the comparison with the conventional thin-bed inversion algorithms shows that the algorithm is not sensitive to the variation of reflection coefficient of reservoir top and bottom interface.Finally,the two-step MP algorithm was applied to spectral decomposition of field data,meanwhile, based on the slices of different frequency,the gas shadow phenomenon below gas-bearing reservoir was analyzed,the results show that shadow phenomenon can directly indicate the existence of hydrocarbon.更多还原