基于二维图像和多点统计方法的多孔介质三维重构研究  

A method of reconstruction of porous media using a two-dimensional image and multiple-point statistics

作　　者：张挺[1,2] 卢德唐[1,2] 李道伦[1,2] 

机构地区：[1]中国科学技术大学近代力学系,安徽合肥230027 [2]中国科学技术大学石油与天然气研究中心,安徽合肥230027

出　　处：《中国科学技术大学学报》2010年第3期271-277,共7页JUSTC

基　　金：国家自然科学重点基金(10932010);国家重大科技专项(2008ZX05001);中科院知识创新重大项目(KJCX1-YW-21);国家自然科学基金(10702069);国家重点基础研究发展(973)计划(2006CB705805)资助

摘　　要：提出了一种基于多孔介质二维薄片图像和多点统计方法的多孔介质三维重构方法.将微米精度的多孔介质二维薄片图像作为初始的训练图像,利用多点统计方法重构其下层图像.将每次获得的下层图像作为新的训练图像,利用多点统计方法继续重构该训练图像的下层图像.二维重构图像的每个像素对应成三维空间的一个体素,依次叠加每层的二维重构图像得到最终的三维多孔介质图像.将该方法应用于砂岩样品的三维重构,通过与真实情况下砂岩体数据的比较发现,该方法重构的多孔介质具有与其相似的孔隙结构特征.A method using a two-dimensional thin section image of porous media and multiple-point statistics to reconstruct the three-dimensional structures of porous media was proposed.A two-dimensional thin section image of porous media with the resolution of micron was used as an original training image.The two-dimensional reconstruction image of the next layer obtained each time was taken as a new training image.Multiple-point statistics is used to reconstruct the image of the next layer of the new training image.Each pixel in the two-dimensional images corresponded to a voxel in the three-dimensional space.The final three-dimensional image of porous media was obtained by stacking all the reconstructed images of each layer successively.This method was tested on the three-dimensional reconstruction of sandstone samples.Comparison with experimental results indicate that the characteristics of pore structures in porous media reconstructed through this method were similar to those of real pore structure in the volume data of sandstone samples.

关 键 词：多点统计方法 多孔介质 重构 训练图像 

分 类 号：P628[天文地球—地质矿产勘探]