Scholarlink

<p>Landslide disasters cause significant economic losses and casualties. Accurately identifying landslide failure modes is crucial for monitoring and early warning. This study proposes a general classification system of four failure modes. Based on 219 landslide cases collected from literature, the study uses slope angle, material, rock stratum structure, and dip angle as prediction indicators. Five machine learning algorithms are applied, with non-numerical indicators processed by one-hot encoding. A parameterization scheme with optimal effects is determined through comparisons. Adjusting neural network parameters shows that the CNN algorithm performs best, but it has limitations in distinguishing between buckling and toppling fracture plane sliding of cataclinal rock slopes due to overlapping data distribution and limited sample size. Overall, the research results landslide failure mode identification, providing a reference for enhancing monitoring and early warning capabilities. It offers practical significance and technical support for landslide disaster prevention and control.</p>

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

A normal classification system and intelligent identification method for slope failure

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月6日

查看全文
  • Dong Li,
  • Jun-yi Jiang,
  • Ren-jie Li,
  • Zhen-hua Zhang,
  • Wen-ru Zhao,
  • Ming-liang Chen,
  • Xun-hao Li,
  • Qun-jian Zhang

摘要

Landslide disasters cause significant economic losses and casualties. Accurately identifying landslide failure modes is crucial for monitoring and early warning. This study proposes a general classification system of four failure modes. Based on 219 landslide cases collected from literature, the study uses slope angle, material, rock stratum structure, and dip angle as prediction indicators. Five machine learning algorithms are applied, with non-numerical indicators processed by one-hot encoding. A parameterization scheme with optimal effects is determined through comparisons. Adjusting neural network parameters shows that the CNN algorithm performs best, but it has limitations in distinguishing between buckling and toppling fracture plane sliding of cataclinal rock slopes due to overlapping data distribution and limited sample size. Overall, the research results landslide failure mode identification, providing a reference for enhancing monitoring and early warning capabilities. It offers practical significance and technical support for landslide disaster prevention and control.