论文详情
系统分析了现有地震数据高分辨率处理方法的适用性及不足, 探索了一种基于高频噪声定量约束函数的高分辨率处理方法, 有效且保真地提高地震资料信噪比与分辨率。通过分析叠后地震数据有效信号带宽, 设计了压制高频噪声的最大概率准则和定量约束函数, 通过多次迭代压制噪声, 提高全频带信号的信噪比, 逐步分级拓宽地震有效频带宽度, 循环迭代求取反射系数, 将提取的地震子波与反射系数褶积得到高分辨率地震数据体。测试校验对比其它高分辨率地震处理结果, 处理后的数据体具有垂向分辨率高、主频高且频带宽的特点, 有效视主频由50 Hz提高至100 Hz左右, 地质目标视分辨率相应提高近1倍。基于该提频数据与波阻抗反演实现了2~5 m地质目标的定性与定量预测, 钻井验证符合率90%以上, 有效指导井位部署。该方法弥补了在传统“有效带宽”内提高地震高频信号的不足, 实现了全频带内重构高、低频有效信号的目的。
Based on the discussion of the applicability and limitations of existing high-resolution processing techniques, we propose a new method based on the quantitative constraint function of high-frequency noises to improve seismic resolution and signal-to-noise ratio in the context of amplitude preservation. In view of the effective bandwidth of poststack seismic data, we formulate the maximum probability criterion and quantitative constraint function for iterative high-frequency noise reduction to expand effective frequency band step by step and improve the signal-to-noise ratio of full-band data. Reflection coefficients calculated iteratively are convolved with seismic wavelet to obtain high-resolution seismic data. Compared with other high-resolution processing techniques, our method yields a data volume with high vertical resolution, high dominant frequency, and wide band; effective apparent dominant frequency increases from 50 to 100 Hz, and apparent resolution nearly doubles. Based on the frequency data and wave impedance inversion, the qualitative and quantitative prediction of 2~5 meters geological targets is realized, and the verfication test matches with more than 90% of the targets, which effectively guides the well location deployment. This new method performs better than routine techniques in high-frequency signal enhancement within the effective frequency band and thus realizes high- and low-frequency signal reconstruction for full-band data.