现代交通技术2025，Vol.22Issue(4)：47-52,6.
基于VGG结合迁移学习算法的隧道围岩分级研究
Research on Tunnel Surrounding Rock Classification Based on VGG Combined with Transfer Learning Algorithm
摘要
Abstract
To enhance the accuracy of surrounding rock classification during tunnel construction,the surrounding rocks of tun-nels along the Jingning-Wencheng section of the Longliwen Expressway in Zhejiang Province are investigated.The distribution of surrounding rock classification and lithological characteristics is analyzed.By employing data processing techniques such as scaling and cropping,the quantity of images in the model training set was expanded.The model structures of the KNN algorithm and the VGG combined with transfer learning algorithm were analyzed.Regression models for directly predic-ting modified quality indices of rock masses and classification models for predicting surrounding rock grades were constructed using these two algorithms,respectively.The results show that the VGG combined with transfer learning algorithm achieves su-perior performance in classifying surrounding rock images.The methodology and outcomes can provide valuable references for surrounding rock classification and support design assistance.关键词
围岩分级/KNN算法/VGG算法/迁移学习Key words
surrounding rock classification/KNN algorithm/VGG algorithm/transfer learning分类
交通工程引用本文复制引用
渠成堃,吴德兴,李伟平,张传庆..基于VGG结合迁移学习算法的隧道围岩分级研究[J].现代交通技术,2025,22(4):47-52,6.基金项目
浙江省交通运输厅科技计划项目(2019010、202213) （2019010、202213）
浙江省基础公益研究计划项目(LGF19E080005) （LGF19E080005）
