首页|Utilizing semantic-level computer vision for fracture trace characterization of hard rock pillars in underground space

Utilizing semantic-level computer vision for fracture trace characterization of hard rock pillars in underground space

扫码查看
原文链接
  • 万方数据
  • 维普
This study utilizes a semantic-level computer vision-based detection to characterize fracture traces of hard rock pillars in underground space.The trace images captured by photogrammetry are used to estab-lish the database for training two convolutional neural network(CNN)-based models,i.e.,U-Net(University of Freiburg,Germany)and DeepLabV3+(Google,USA)models.Chain code technology,poly-line approximation algorithm,and the circular window scanning approach are combined to quantify the main characteristics of fracture traces on flat and uneven surfaces,including trace length,dip angle,den-sity,and intensity.The extraction results indicate that the CNN-based models have better performances than the edge detection methods-based Canny and Sobel operators for extracting the trace and reducing noise,especially the DeepLabV3+model.Furthermore,the quantization results further prove the reliabil-ity of extracting the fracture trace.As a result,a case study with two types of traces(i.e.,on flat and uneven surfaces)demonstrates that the applied semantic-level computer vision detection is an accurate and efficient approach for characterizing the fracture trace of hard rock pillars.

Fracture traceHard rock pillarSemantic-level computer visionConvolutional neural network(CNN)Underground space

Chuanqi Li、Jian Zhou、Daniel Dias

展开 >

Laboratory 3SR,CNRS UMR 5521,Grenoble Alpes University,Grenoble 38000,France

School of Resources and Safety Engineering,Central South University,Changsha 410083,China

国家自然科学基金Outstanding Youth Project of Hunan Provincial Department of EducationDistinguished Youth Science Foundation of Hunan Province of China国家留学基金委项目Sinosteel Maanshan General Institute of Mining Research Co.,Ltd for providing data acquisition equipment

4217716423B00082022JJ10073202106370038Conon EOS-5D Mark Ⅲ

2024

10.1016/j.gsf.2023.101769

地学前缘(英文版)
中国地质大学（北京） 北京大学

地学前缘(英文版)

CSTPCD
影响因子：0.576
ISSN：1674-9871
年,卷(期)：2024.15(2)
  • 100