Scholarlink

<p>An increasingly important field of research has emerged in recent decades: the interaction between urban infrastructure, human lives, and natural disasters. Analysis and reaction to natural disasters have made extensive use of deep learning (DL) techniques using semantic segmentation networks. The aim of this research is to develop novel method in disaster-based geographical region changes and climate change detection utilising DL model. Here, input is collected as disaster-based geographical region and climate change dataset as well as processed for noise removal and smoothening. Then, this image feature has been extracted and classified using watershed U-net component analysis with Hopfield variational fuzzy encoder neural networks. The fundamental component of the suggested approach is a deep learning procedure that makes it possible to identify soccer fields and analyse helicopter landing sites. The technique uses satellite photos and provided geographic data to train a deep learning autoencoder. Experimental analysis has been carried out in terms of classification accuracy, specificity, sensitivity, ROC, and <i>F</i>-measure. The proposed technique obtained 96% of classification accuracy, 94% of sensitivity, 98% of specificity, 95% of ROC, and 97% of <i>F</i>-measure.</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

Disaster-Based Geographical Region Analysis Using Climate Change Detection Using Deep Learning Algorithm

  • 期刊论文

  • 发表日期：

    2025年4月17日

查看全文
  • P. K. Swapna,
  • Rahul Ganpat Mapari,
  • Elangovan Muniyandy,
  • Jacquline Tham,
  • M. Sandra Carmel Sophia,
  • Ritwik Haldar

摘要

An increasingly important field of research has emerged in recent decades: the interaction between urban infrastructure, human lives, and natural disasters. Analysis and reaction to natural disasters have made extensive use of deep learning (DL) techniques using semantic segmentation networks. The aim of this research is to develop novel method in disaster-based geographical region changes and climate change detection utilising DL model. Here, input is collected as disaster-based geographical region and climate change dataset as well as processed for noise removal and smoothening. Then, this image feature has been extracted and classified using watershed U-net component analysis with Hopfield variational fuzzy encoder neural networks. The fundamental component of the suggested approach is a deep learning procedure that makes it possible to identify soccer fields and analyse helicopter landing sites. The technique uses satellite photos and provided geographic data to train a deep learning autoencoder. Experimental analysis has been carried out in terms of classification accuracy, specificity, sensitivity, ROC, and F-measure. The proposed technique obtained 96% of classification accuracy, 94% of sensitivity, 98% of specificity, 95% of ROC, and 97% of F-measure.