适用于超大规模三维地震层位自动追踪的双尺度洪水填充方法

Double-scale flood fill approach for large-scale 3D horizon auto-tracking

双尺度洪水填充是一种适用于超大规模三维地震层位空间自动追踪的方法。该方法从三个方面对经典洪水填充(Flood fill)算法进行扩展,即首先根据简单分块算法对三维地震工区进行分块并建立两级队列;再根据种子点顺序进行区块编号并实施区块级洪水填充;最后基于镶边分块算法对区块进行扩边并以区块为单位实施种子点级洪水填充。改进后的双尺度洪水填充法用"区块"和"种子点"两种尺度队列代替经典洪水填充法中的"种子点"一种队列,采用一致的四连通洪水填充法实现种子点扩散,在确保层位追踪的准确性和精度的前提下,显著降低了大规模或超大规模三维层位自动追踪对计算机资源的需求。

Double-scale flood fill is an innovative spatial horizon auto-tracking technique suitable for largescale 3D seismic data.It is improved from classical flood fill for seismic horizon auto-tracking and it borrows some benefits from block partition algorithm.The significant advantage of the improved method is that it uses two-level queues-block and seed queues-instead of classical one-level queue and the same four-way flood fill mechanism to implement horizon auto-tracking.In this paper,we first discuss the theory of double-scale flood fill algorithm,then introduce a typical workflow which can be used to implement spacious horizon auto-tracking,and finally give a real 3D horizon auto-tracking example.The satisfactory quality of horizon autotracking proves that the proposed double-scale flood fill approach is not only meet the requirements of accurate horizon auto-tracking under the condition obeying the rule of the classical flood fill,but also can remarkably reduce computer resource consumption and improve horizon-tracking efficiency in large-scale or ultra large-scale 3D seismic data volume.