Scholarlink

<p>Intense rainfall-induced landslides provide significant dangers to human habitation and infrastructure in hilly regions. The precise and prompt identification continues to pose a significant problem due to the intricate interaction of geological and environmental elements. This work presents a Transformer-based Landslide Detection model (T-LDet) for the real-time prediction of rainfall-induced slope failure. The proposed model employs the transformer architecture to efficiently capture temporal correlations among rainfall, soil moisture, hydrogeological, and landscape characteristics. T-LDet, applied to data from the active landslide zone of Mawiongrim, Meghalaya, India, demonstrates superior predicted accuracy and reduced error metrics compared to traditional machine learning models. The findings confirm T-LDet as a dependable foundation for real-time landslide prediction and early warning systems.</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

T-LDet: enhancing landslide detection using transformer-driven approach in Mawiongrim, Meghalaya, India

  • 期刊论文

  • 发表日期：

    2025年12月20日

查看全文
  • Namrata Govind Ambekar,
  • Surmila Thokchom,
  • Smrutirekha Sahoo

摘要

Intense rainfall-induced landslides provide significant dangers to human habitation and infrastructure in hilly regions. The precise and prompt identification continues to pose a significant problem due to the intricate interaction of geological and environmental elements. This work presents a Transformer-based Landslide Detection model (T-LDet) for the real-time prediction of rainfall-induced slope failure. The proposed model employs the transformer architecture to efficiently capture temporal correlations among rainfall, soil moisture, hydrogeological, and landscape characteristics. T-LDet, applied to data from the active landslide zone of Mawiongrim, Meghalaya, India, demonstrates superior predicted accuracy and reduced error metrics compared to traditional machine learning models. The findings confirm T-LDet as a dependable foundation for real-time landslide prediction and early warning systems.