煤田地质与勘探2025，Vol.53Issue(2)：55-66,12.DOI:10.12363/issn.1001-1986.24.09.0609
基于μCT和深度学习的煤裂隙智能提取方法
A method for intelligent information extraction of coal fractures based on μCT and deep learning
摘要
Abstract
[Objective]The fine-scale characterization of fractures in coal reservoirs is significant for the exploration and exploitation of coalbed methane(CBM)resources.Given that the size,orientation,and density of fractures directly af-fect the permeability of coal seams,the accurate information identification and extraction of fractures in coal seams plays a key role in revealing the formation and propagation mechanisms of fracture networks during reservoir volume fractur-ing.Conventional methods for fracture information extraction typically rely on manual labeling and feature extraction based on image processing techniques,exhibiting significantly limited accuracy and efficiency.[Methods]This study proposed a method for fracture information extraction of coals based on TransUNet and micro-computed tomography(μCT)images.TransUNet,integrating the advantages of both the Transformer modules and convolutional neural net-work(CNN),is capable of extracting global features and capturing local details in images,significantly enhancing the image segmentation accuracy and network robustness.First,the μCT images of coal samples were preprocessed,includ-ing improving the image quality using the difference method and increasing the sample size using data augmentation techniques.Subsequently,image segmentation was conducted using TransUNet to extract fracture features.Additionally,the image segmentation results of varying neural network models were compared.[Results and Conclusions]The res-ults indicate that the proposed method exhibited superior performance on a given dataset.Specifically,the TransUNet model yielded an accuracy of 91.3%,precision of 89.5%,F1 score of 89.8%,and Intersection over Union(IoU)of 84.0%,significantly outperforming other intelligent models like U-Net and U-Net++.Given the characteristics of fine-grained μCT images,applying TransUNet to the fracture information extraction of coals emerges as an efficient and ac-curate approach.This study provides a novel philosophy for image processing in the field of CBM exploration and pro-duction.关键词
Trans-UNet/μCT图像/煤裂隙/图像分割/深度学习Key words
Trans-UNet/μCT images/coal fracture/image segmentation/deep learning分类
地球科学引用本文复制引用
胡咤咤,张寻,金毅,巩林贤,黄文辉,任建吉,Norbert Klitzsch..基于μCT和深度学习的煤裂隙智能提取方法[J].煤田地质与勘探,2025,53(2):55-66,12.基金项目
国家自然科学基金青年项目(42402184) （42402184）
河南省科技攻关项目(232102320200) （232102320200）
河南省高等学校重点科研项目(23A170006) （23A170006）
