基于非常快速模拟退火算法的地震谱反演方法

Spectral Inversion Based on Very Fast Simulated Annealing Algorithm

摘要:

地震勘探中地层薄层特征的描述对提高地震分辨率十分重要。谱反演是建立在谱分解、时频分析基础上的一种反演方法。本文在已知子波的情况下,采用全局搜索的非常快速模拟退火算法(VFSA)对高信噪比的叠后地震资料进行谱反演。有效的识别了调谐厚度以下的薄层,得到近似反射系数的剖面,同时拓展了原始资料的频带宽度,提高了资料的分辨率,为解释资料的局部细微特征起到重要作用。

Abstract:

The description of thin bed in seismic prospecting is very important for obtaining high frequency.Spectral inversion is a method based on the spectral resolution and time-frequency analysis,which can effectively determine the layer thickness below the seismic tuning thickness.In this paper,a global search method——Very Fast Simulated Annealing（VFSA） is used to spectral inversion in high SNR post-stack seismic data.As the presence of noise,in experiment,band spectrum is used in spectral inversion.The result is Quiasi-reflection coefficient.The method expands the bandwidth of the original data,and enhances the data resolution.It is an important role to explain the subtle characteristics in seismic prospecting.