本文已被：浏览 1144次 下载 432次
投稿时间：2019-04-08
投稿时间：2019-04-08
中文摘要: 岩石的内部结构复杂，虽然利用扫描电镜可获得精确的高分辨率页岩孔隙结构，但是实验成本高，耗时长，不便于大规模运用。针对这一问题，提出了基于无监督卷积神经网络的页岩重构方法，结合页岩图像软数据，进行了页岩重构。实验证明该方法只需要少量的真实页岩数据即可获得较好的重构结果；与经典的数值重构方法Snesim和Filtersim方法相比，该方法耗时更少，具有一定优势。
Abstract:The internal structure of the rock is complex.Although accurate high-resolution shale pore structure can be obtained by scanning electron microscopy,the experiment is expensive,time-consuming and inconvenient for large-scale application.A shale reconstruction method based on unsupervised convolutional neural network is proposed,which combines shale image soft data to reconstruct shale.This method only needs a small amount of real shale data to obtain better reconstruction results.Compared with the Snesim and Filtersim methods for shale reconstruction,this method has advantages.
keywords: reconstruction shale convolutional neural network
文章编号：20204008 中图分类号：TP399 文献标志码：
基金项目:国家自然科学基金（41672114，41702148）。
引用文本：
张挺,张瑜,杜奕.基于卷积神经网络的页岩重构方法[J].上海电力大学学报,2020,36(4):364-368.
ZHANG Ting,ZHANG Yu,DU Yi.A Shale Reconstruction Method Based on Convolutional Neural Networks[J].Journal of Shanghai University of Electric Power,2020,36(4):364-368.
张挺,张瑜,杜奕.基于卷积神经网络的页岩重构方法[J].上海电力大学学报,2020,36(4):364-368.
ZHANG Ting,ZHANG Yu,DU Yi.A Shale Reconstruction Method Based on Convolutional Neural Networks[J].Journal of Shanghai University of Electric Power,2020,36(4):364-368.