Scholarlink

<p>Landslides are abrupt geological events posing significant hazards to human safety and environmental stability. This review provides an extensive analysis of the latest methodologies in landslide susceptibility prediction, from traditional statistical models to advanced machine learning and deep learning techniques. Emphasizing the importance of integrating diverse datasets—including remote sensing imagery, geological maps, and meteorological data—this paper categorizes existing research into various approaches such as dataset preparation, contributing factors, feature extraction tools, and methodological frameworks along with discussions on new cloud-based techniques as well as the impact of climate change on landslide susceptibility prediction using CMIP6 and RCP. The review also emphasizes the validity of the previously used algorithms in tackling the issue at hand. It highlights the increasing role of interdisciplinary techniques in enhancing predictive accuracy and addresses notable research gaps. To bridge these gaps, the review proposes a novel one-class classifier (OCSVM) for landslide prediction, utilizing a custom dataset derived from user-provided geo-coordinates and dynamic extraction methods from tools like Google Earth and GIS, along with real-time weather data. The findings highlight the significance of continuous innovation in predictive models to improve landslide risk management and early warning systems, ultimately aiming to safeguard communities from these natural threats.</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

Datasets, Features, and Advanced Techniques in Landslide Susceptibility Prediction: a Review

  • 期刊论文

  • 发表日期：

    2025年9月26日

查看全文
  • Sarvesh Kulkarni,
  • Samyak Jasani,
  • Sarang Kulkarni,
  • Jyoti Wadmare

摘要

Landslides are abrupt geological events posing significant hazards to human safety and environmental stability. This review provides an extensive analysis of the latest methodologies in landslide susceptibility prediction, from traditional statistical models to advanced machine learning and deep learning techniques. Emphasizing the importance of integrating diverse datasets—including remote sensing imagery, geological maps, and meteorological data—this paper categorizes existing research into various approaches such as dataset preparation, contributing factors, feature extraction tools, and methodological frameworks along with discussions on new cloud-based techniques as well as the impact of climate change on landslide susceptibility prediction using CMIP6 and RCP. The review also emphasizes the validity of the previously used algorithms in tackling the issue at hand. It highlights the increasing role of interdisciplinary techniques in enhancing predictive accuracy and addresses notable research gaps. To bridge these gaps, the review proposes a novel one-class classifier (OCSVM) for landslide prediction, utilizing a custom dataset derived from user-provided geo-coordinates and dynamic extraction methods from tools like Google Earth and GIS, along with real-time weather data. The findings highlight the significance of continuous innovation in predictive models to improve landslide risk management and early warning systems, ultimately aiming to safeguard communities from these natural threats.