基于深度学习的裂隙智能提取研究

打开文本图片集

中图分类号：TP18 文献标志码：A 文章编号：2095-2945（2025）13-0009-05

Abstract：WiththerapiddevelopmentofChina’seconomicconstruction，engineeringconstructionismovinginahigher， deperandbroaderdirection.Inordertoensurethesafetyof projectconstruction，itisnecessarytoidentifytheenginering geologicalconditionsofthesitebeforeprojectconstruction.Asanimportantpartofrockmassstructure，fractureshavean importantimpactonengineeringgeologicalconditions.Therefore，extractingthestructureofcracksisveryimportantfor engineeringconstruction.Traditionalcrack structureextractionmethodsaretime-consumingandlabor-intensive，andhavepoor operability;theacuracyofcrackextractionmethodsusingtraditionalcomputertechologycannotmeetengineeringneedsandare poorinpracticality;however，therearefewresearchonusingdeeplearningtechnologytoextractfractures.Bystudyingthedeep learningnetworkoftheencoder-decoderarchitecture，thispaperdeterminesthattheimagesegmentationmodeltrainedbythe "U-Net++"frameworknetworkandthe"Resnet5O"encoder-decodernetworkcanefectivelyextractthefisurestructureoffield outcrop photos.

Keywords： crack extraction; image segmentation； deep learning; crack structure; extraction method

随着我国经济建设的高速发展，房屋建筑越来越高，地基基础越来越深，穿山隧道越来越长，水利水电工程越来越庞大，相应地，对于工程建设过程中的工程地质条件要求也越来越高，对工程地质信息的掌握也要越来越精确。（剩余6462字）