非线性各向异性扩散滤波器用于三维地震资料噪声衰减与结构特征增强

Non-linear anisotropic diffusion filtering for 3D seismic noise removal and structure enhancement

消除地震资料中的干扰噪声,增强地质沉积特征,对地震资料的后继处理和解释都有重要的意义.本文重点论述一种基于偏微分方程的非线性各向异性扩散滤波技术.采用梯度结构张量估计局部地层方向,度量地质特征的规则程度,以此约束扩散滤波器沿着地层方向作扩散滤波,并且通过控制在断层和河道等不连续性区域的滤波程度,使得扩散滤波器具有优异的保持边缘特性.对大庆油田的一块实际地震资料的处理结果表明,该方法可有效消除非相干噪声,增强地震反射同相轴的连续性,突出地质沉积特征,提高资料的信噪比,整体改善地震资料的品质.

It makes important sense to remove random seismic noise and enhance geologic features, which can ease subsequent processing and interpretation of 3D seismic data. This paper focuses on discussing a partial differential equation based nonlinear anisotropic diffusion filtering technique. We use the Gradient Structure Tensor to estimate local orientations of seismic events and measure the regularity of local seismic structures. So that, the diffusion process is just implemented along seismic events and the amount of smoothing is controlled in discontinuous regions such as faults and channels. This makes the diffusion filter perform excellent edge preserving smoothing. The filter wias applied to a 3D seismic data from Daqing Oilfield and the filtering results demonstrate its efficiency to remove incoherent noise as well as to enhance the continuity of seismic events and geologic sediments. This diffusion filter can improve the quality of seismic data wholly.