地震全波形反演(FWI)主要利用低频长偏移距透射波估计背景速度,有效的初始背景速度场对于透射波FWI和反射波FWI的有效应用都十分重要.因此,根据散乱的速度和界面深度信息,对不同来源、不同精度的散乱速度场插值产生背景速度模型是速度建模的核心环节之一.在总结了三角剖分法、自然邻域法、反距离加权插值法和径向基函数插值法以及克里金插值法等方法特点的基础上,采用三维克里金插值方法将实际测井数据插值成三维速度场,然后利用克里金方差融合策略和多尺度空间波数域Gabor变换两种方法对分别来自测井数据和来自偏移速度分析的速度场进行融合,融合后的速度场的精度得到了提高,成像质量明显改善.最后,对基于散乱数据支撑的特征速度场的描述问题进行了讨论,提出了多尺度特征速度场的表达策略.%Full waveform inversion is the current hot research topic in the field of seismic exploration,mainly using the low frequency and long offset transmitted wave to estimate the background velocity.The lack of effective low frequency component about seismic data is the dominating problem in land seismic exploration.The accuracy and validity of initial background velocity is very important for transmission FWI and reflection wave FWI.According to the scattered velocity data and interface depth information,no matter where they come from,interpolating the scattered data to produce an accurate background velocity model is the core of velocity model building.We firstly summarize some of the scattered data interpolation methods such as triangulation based interpolation,natural neighbor interpolation,inverse distance weighted,radial basis function interpolation and Kriging interpolation.Then,we employ 3D Kriging interpolation method to interpolate the practical log data into 3D velocity field and provide two methods to merge the velocity fields from log data and velocity analysis for migration.One of these methods is utilizing Kriging variance and the other one is making use of the multiscale Gabor transform.After that we apply the merged velocity fields to migration imaging process.Application results show that the accuracy of velocity field can be enhanced and both these merged methods improve the imaging quality.In the end,we discuss the description problem on the characteristics of the velocity field based on scattered data interpolation and put forward the multi-scale expression of characteristic velocity field.
展开▼
