Scholarlink

In today’s era of rapid urbanization and environmental challenges, effective disaster and crisis management demand innovative solutions. This paper presents a novel approach focusing on community-level water-related issues through an intelligently designed application. The primary objective is to develop a user-friendly platform facilitating the reporting of water-related problems by both social media posts and the general public, subsequently enabling prompt action by relevant government authorities. Leveraging deep learning techniques and user-generated data, our solution introduces real-time detection and classification of six distinct water-related problems. A custom dataset is curated to train a ResNet-18 model, achieving an impressive accuracy of 73%. The application, developed with a React-based frontend and Flask-powered backend, acts as a centralized hub for reporting and managing water issues. Notably, it employs user-inputted data to accurately pinpoint problem locations, thereby enhancing the precision of reporting. By presenting a holistic approach, this research significantly contributes to the development of efficient crisis management and response strategies for water-related disasters.

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

AquaMap: Empowering Communities to Report and Map Water-Related Issues in Real-Time with Deep Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Harshitha Lakshmi Durga Nalla,
  • Anusha Bhuchupalli,
  • Tejasree Addala,
  • Yasasri Sabbineni,
  • Koppisetti Sravya Geetha,
  • Ghantasala Aasha,
  • Sridevi Bonthu

摘要

In today’s era of rapid urbanization and environmental challenges, effective disaster and crisis management demand innovative solutions. This paper presents a novel approach focusing on community-level water-related issues through an intelligently designed application. The primary objective is to develop a user-friendly platform facilitating the reporting of water-related problems by both social media posts and the general public, subsequently enabling prompt action by relevant government authorities. Leveraging deep learning techniques and user-generated data, our solution introduces real-time detection and classification of six distinct water-related problems. A custom dataset is curated to train a ResNet-18 model, achieving an impressive accuracy of 73%. The application, developed with a React-based frontend and Flask-powered backend, acts as a centralized hub for reporting and managing water issues. Notably, it employs user-inputted data to accurately pinpoint problem locations, thereby enhancing the precision of reporting. By presenting a holistic approach, this research significantly contributes to the development of efficient crisis management and response strategies for water-related disasters.