改进的各向异性扩散滤波方法压制地震数据噪声

Seismic data noise attenuation by improved nonlinear anisotropic diffusion filtering method

闫安菊¹ 蔡涵鹏²

成都理工大学地球物理学院，四川 成都 610059 中国石油川庆钻探工程有限公司地球物理勘探公司，四川 成都 610213

College of Geophysics, Chengdu University of Technology, Chengdu 610059, China Geophysical Exploration Company, Chuanqing Drilling Engineering Co. Ltd., CNPC, Chengdu 610213, China

在压制地震数据噪声时，为有效保持地质体边界、断层等细节特征信息，根据地震数据噪声多尺度特性，分析已有的各向异性扩散滤波模型的特点，将控制扩散特性的梯度阀值、迭代时间和高斯滤波尺度因子参数引入扩散模型中，并借鉴模拟退火的冷却思想，设定高斯滤波尺度因子和梯度阀值为时间的函数，构建了能够保持边缘特征、压制地震数据噪声的各向异性扩散滤波模型。数值试算和实际资料的应用表明，构建的各向异性扩散滤波模型能够自适应地、有效地压制地震数据噪声，为地震数据解释和分析、属性提取提供了较好的基础数据。

In order to solve the problems of preserving the characteristics of geological boundary and fault when attenuating the seismic data noise and according to the multi-scale characteristics of seismic data noise, the features of existing nonlinear anisotropic diffusion filtering model are analyzed. The gradient threshold value for controlling the diffusion properties, iterative time and Gauss filter scaling factor parameters are introduced into the diffusion model. The Gauss filter scaling factor and the gradient threshold value are set as a function of time in accordance with simulated annealing idea. The anisotropic diffusion filtering model which can preserve the edge characteristics and attenuating the seismic data noise is built. Numerical calculation and the application of actual data show that the improved anisotropic diffusion filtering model can self-adaptively and effectively attenuating noise, which can provide a good basic data for seismic data interpretation, analysis and attribute extraction.

10.6056/dkyqt201205011