Scholarlink

Accurate identification of rock joints is critically important for ensuring the safety and stability of slope engineering projects. By effectively predicting potential risks of rock mass sliding and collapse, it allows for the optimization of design and construction plans, thereby significantly enhancing the overall safety of engineering projects. In this study, we present an intelligent rock joint recognition model, TransUNet, which combines the architectures of Transformer and UNet. This model leverages the self-attention mechanism of the Vision Transformer (ViT) to process global information while capturing local spatial details through UNet, resulting in a notable improvement in recognition accuracy and efficiency. In the Akurenam-Minang highway project, TransUNet demonstrated outstanding rock joint recognition capabilities, showcasing the immense potential of modern computer vision and deep learning technologies in this field. The model provides reliable technical support for the identification and analysis of rock in slope engineering and contributes to the advancement of intelligent rock joint recognition technology.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Research on Rock Joint Recognition Based on Transformer-UNet Deep Learning Method for Engineering Disaster Prevention

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Cong Chen,
  • Wenbo Wang,
  • Jiefeng Zhang,
  • Huixing Li,
  • Zicheng Li,
  • Jing Zhang

摘要

Accurate identification of rock joints is critically important for ensuring the safety and stability of slope engineering projects. By effectively predicting potential risks of rock mass sliding and collapse, it allows for the optimization of design and construction plans, thereby significantly enhancing the overall safety of engineering projects. In this study, we present an intelligent rock joint recognition model, TransUNet, which combines the architectures of Transformer and UNet. This model leverages the self-attention mechanism of the Vision Transformer (ViT) to process global information while capturing local spatial details through UNet, resulting in a notable improvement in recognition accuracy and efficiency. In the Akurenam-Minang highway project, TransUNet demonstrated outstanding rock joint recognition capabilities, showcasing the immense potential of modern computer vision and deep learning technologies in this field. The model provides reliable technical support for the identification and analysis of rock in slope engineering and contributes to the advancement of intelligent rock joint recognition technology.