基于孔隙度分类的超维重建算法

Super dimension reconstruction algorithm based on porosity classification

摘要:

由于超维算法在字典建立过程和三维重建过程都涉及到大量的模式匹配，导致超维算法耗时较长，在实际的应用中还有一定难度.针对这个问题，本文提出了一种基于孔隙度分类的超维重建算法，能够较大程度上减少三维重建的时间成本.首先，结合字典元素的孔隙度这一特征对字典集进行分类；其次，利用孔隙度分类字典在重建时采取依孔隙度的搜索方式，优先搜索相应的字典区间；然后，针对不同的训练图像进行三维重建，结合孔隙度分布，提出了一种自适应的搜索范围确定方法；最后，通过对高中低三种孔隙度的训练图像分别进行多次重建，将传统超维算法和新算法的重建结果与真实岩心三维结构的统计特征函数、孔隙参数以及两种算法重建时间的进行对比分析，验证算法的有效性.

Abstract:

Since a lot of pattern matching in the process of dictionary set establishment and three-dimensional(3D) reconstruction, the super dimension algorithm takes a long time and is difficult in practical application. To solve this problem, a super-dimensional reconstruction algorithm based on porosity classification is proposed, which can greatly reduce the time cost of 3D reconstruction. First, the dictionary sets are classified by the porosity of dictionary elements. Then, the porosity classification dictionary is used to search according to porosity during reconstruction, and the corresponding dictionary interval is searched first. Based on the 3D reconstruction of different training images and porosity distribution, an adaptive search range determination method is proposed. Finally, the effectiveness of the proposed super dimension algorithm is verified by multiple reconstruction of the training images of high, medium and low porosity, a comparative analysis is made on the reconstruction results of the traditional Super dimension algorithm and the proposed algorithm in terms of the statistical characteristic function, pore-throat parameters and reconstruction time.