基于CT图像的孔隙结构参数提取算法改进

摘要

Quantitative characterization of pore network model extracting microscopic pore space is the newest ways of quantitative characterization of rock three-dimensional complex pore structure.The usual method is the largest ball algorithm.But the length of the hole is too large and the throat length is too small.That would influence the determination of pore and throat parameters.For this problem,two key algorithms for geometric transformation and discriminant analysis was introduced in this paper.And it achieves a reasonable division of the pores and throats and pore size distribution.Compared with rock pore size distribution from nuclear magnetic T2 spectrum conversion,the improved maximum ball algorithm can accurately extract the pore size distribution of the rock.%对岩石三维复杂孔隙结构进行定量表征最新的方法是基于三维数字岩心,通过提取微观孔隙空间的孔隙网络模型进行定量表征.目前常用的方法是最大球算法,但该算法确定的孔隙长度偏大而喉道长度偏小,间接影响其他孔隙和喉道参数的确定.针对遇到的问题,在最大球算法的基础上通过引入图形几何变换和判别分析2项关键算法,实现了孔隙、喉道的合理划分,准确提取岩石的孔径分布.与核磁共振T2谱转换的岩石孔径分布对比表明,改进的最大球算法能够准确地提取岩石的孔径分布.