基于非平稳相似性系数的构造导向滤波及断层检测方法

不连续地质体（如断层）的自动检测一直以来都是叠后地震数据解释中的关键问题之一，尤其在三维情况中尤为重要.然而，大多数边缘检测和相干算法都对随机噪声很敏感，随机噪声衰减是叠后地震数据解释的另一个主要问题.针对构造保护去噪和断层检测问题，本文基于非平稳相似性系数完善一种构造导向滤波方法并且提出一种自动断层检测方法，形成了一套匹配的处理技术.该构造导向滤波既能够有效地衰减随机噪声又可以很好地保护地震资料中的断层等信息不被破坏，增强地震剖面中弯曲、倾斜同相轴的连续性.根据地震数据局部倾角走向，利用相邻道构建当前地震道的预测，通过预测道的叠加得到参考道，计算预测道与参考道之间的非平稳相似性系数可以设计出数据驱动的加权中值滤波.另一方面，预测道与原始道之间的非平稳相似性系数能够用于带有断层指示性的相干分析.这两种方法都基于构造预测和非平稳相似性系数，但是使用不同的调节参数和处理方案.理论模型和实际数据的处理结果证明了本文提出构造导向滤波和断层检测方法的有效性.

Structure-oriented filtering and fault detection based on nonstationary similarity

Automatic detection of geological discontinuities, e.g., faults, is a crucial problem especially in 3D seismic interpretation. However, most edge detection and coherence algorithms are often sensitive to random noise. Random noise attenuation is another persistent problem in seismic interpretation. Aiming at both structurally filtering and fault detection, we present a new structure-oriented filter, which aims at reducing random noise while protecting structural information and a new automatic fault detection technique based on nonstationary similarity. We use local dip pattern to form a structure prediction of seismic traces from neighboring traces, the predicted traces then are stacked to be the reference traces. The nonstationary similarity between prediction traces and reference traces can be calculated to design a data-driven weighted median filter, meanwhile, another nonstationary similarity measurement between original data and their prediction data enhances the coherency-type analysis with identification. Both methods are based on structure prediction and nonstationary similarity but with different parameters and strategies. Numerical tests with synthetic and field data confirm the effectiveness of the proposed techniques for structure-oriented filtering and fault detection.