We develop an algorithm for seismic data filtering to preserve the boundary information of faults and other geological bodies while suppressing the noise. Based on the theory of anisotropic diffusion filtering, this method explores the relationship between eigenvalues of the structure tensor and local structural features of three‐dimensional seismic images and innovatively introduces the definition of stratigraphic coherence coefficients. Then we propose a new system to design the eigenvalues of the diffusion tensor by using the fault confidence measure and stratigraphic coherence coefficients, which can control the filtering intensity of seismic data in different orientations. The value of fault confidence measure is close to 0 where there are flat continuous reflectors and the intensity of diffusion is strong. On the contrary, the diffusion is very weak within presumptive fault zones. Results of both synthetic model and real data prove that the proposed method can effectively suppress noise, preserve faults well and enhance the continuity of the reflectors, which can provide basic data with high signal‐to‐noise ratio for subsequent seismic interpretation.

中文翻译：

---

基于断层置信度和地层相干系数的各向异性扩散滤波

我们开发了一种地震数据过滤算法，可在抑制噪声的同时保留断层和其他地质体的边界信息。该方法基于各向异性扩散滤波理论，探讨了结构张量特征值与三维地震图像局部结构特征之间的关系，并创新性地介绍了地层相干系数的定义。然后我们提出了一个新的系统，利用断层置信度和地层相干系数设计扩散张量的特征值，可以控制不同方向地震数据的滤波强度。断层置信度的值接近于0，这时有平坦的连续反射器并且扩散强度很强。相反，推测的断层带内的扩散非常弱。综合模型和真实数据的结果证明，该方法可以有效地抑制噪声，很好地保留断层，增强反射器的连续性，可以为后续地震解释提供高信噪比的基础数据。