利用结构导向平滑方法解释断层

Fault interpretation based on structure-oriented smoothing method

在深层地震资料解释过程中,受资料品质的限制,断层解释难度较大。为此,研究了对叠后地震资料进行处理的基于梯度结构张量的结构导向平滑方法,该方法能有效改善地震资料的品质,进而辅助断层解释。具体流程为:1利用原始地震数据求取每个采样点在空间(或时间)方向的导数,并求取梯度结构张量矩阵;2通过特征分解求取梯度结构张量矩阵的特征向量,并构建扩散张量矩阵;3将扩散张量矩阵代入改进的偏微分扩散方程,采用共轭梯度迭代法求解该方程,求得的解即为平滑后的图像。实际资料处理结果表明:经结构导向平滑方法处理后的地震资料信噪比有较大提高,剖面上的断点更清晰,断层的空间组合关系更明确,解释结果的可靠性更高。

Deep fault interpretation is usually very difficult due to the limitation of seismic data quality.To solve this problem,we try to use the structureoriented smoothing method based on gradient structure tensor for poststack seismic data.The method can effectively improve seismic data quality and fault interpretation.The flowchart is shown as below:(1) Based on raw seismic data,calculate space(or time)direction derivative for each sample,and calculate the gradient structure tensor matrix;(2) Calculate eigenvectors of the gradient structure tensor matrix with eigen-decomposition,and construct the diffusion tensor matrix;(3)Put the diffusion tensor matrix into the improved partial differential diffusion equation and solve the equation with conjugate-gradient iteration method,and the solution represents output smoothed image.Based on our real data applications,seismic data S/N ratio is greatly improved after structure-ori-ented smoothing,and fault breakpoints on seismic sections are clearer.So the interpretation results are more reliable.