Scholarlink

<p>Destructive natural disasters and landslides present serious risks to people’s lives and property. Accurate mapping is challenging because of the complex interplay of causative factors and heterogeneous datasets, while traditional methods often lack effective optimization for improved detection. To overcome these limitations, a Rainfall-Induced Landslide Hazard Mapping approach based on a Multimodal Contrastive Domain-Sharing Generative Adversarial Network (RIL-HM-MCDSGAN) is proposed. At first, the input data are gathered from the landslide causative factor dataset. The data are then subjected to pre-processing, where the Orthogonal Master–Slave Adaptive Notch Filter (OMSANF) is utilized for normalization and handling missing values. The pre-processed output is subsequently fed into feature extraction using the Adaptive Synchro-Extracting Transform (ASET) to extract topographic features. The extracted features are then given to the Multimodal Contrastive Domain-Sharing Generative Adversarial Network (MCDSGAN) for landslide detection and classification into landslide and no landslide categories. Sea-Horse Optimization Algorithm (SHO) is employed to optimize the weight parameters of MCDSGAN, thereby improving the accuracy of landslide detection. The proposed RIL-HM-MCDSGAN approach attains a higher 99.41% accuracy when compared with existing methods.</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

Advanced generative adversarial network for rainfall-induced landslide hazard mapping

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • H. N. Ramya,
  • M. A. Nagesh

摘要

Destructive natural disasters and landslides present serious risks to people’s lives and property. Accurate mapping is challenging because of the complex interplay of causative factors and heterogeneous datasets, while traditional methods often lack effective optimization for improved detection. To overcome these limitations, a Rainfall-Induced Landslide Hazard Mapping approach based on a Multimodal Contrastive Domain-Sharing Generative Adversarial Network (RIL-HM-MCDSGAN) is proposed. At first, the input data are gathered from the landslide causative factor dataset. The data are then subjected to pre-processing, where the Orthogonal Master–Slave Adaptive Notch Filter (OMSANF) is utilized for normalization and handling missing values. The pre-processed output is subsequently fed into feature extraction using the Adaptive Synchro-Extracting Transform (ASET) to extract topographic features. The extracted features are then given to the Multimodal Contrastive Domain-Sharing Generative Adversarial Network (MCDSGAN) for landslide detection and classification into landslide and no landslide categories. Sea-Horse Optimization Algorithm (SHO) is employed to optimize the weight parameters of MCDSGAN, thereby improving the accuracy of landslide detection. The proposed RIL-HM-MCDSGAN approach attains a higher 99.41% accuracy when compared with existing methods.