Scholarlink

Effective management of urban drainage networks is vital for resilience and infrastructure reliability. This paper introduces AquaNetCT, an automated framework for detecting, analyzing, and visualizing manholes in Can Tho city, Vietnam. The system employs computer vision for manhole detection, extracting geospatial and textual metadata based on OCR. Deep learning models (i.e., CNN-DenseNet121) classifies manhole types with over \(95\%\) F1-Score accuracy, enabling prioritized maintenance. The dataset extracted from images/frames is structured and stored into formats (i.e., JSON, CSV, \(\ldots \) ), and depicted using a graph-based algorithm designed to model the drainage network. This framework streamlines inspection workflows and enhances the operational efficiency and sustainability of smart city drainage management.

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

AquaNetCT: Detecting and Representing Urban Drainage Networks from Video Analysis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Thanh Ma,
  • Salem Benferhat,
  • Minh-Thu Tran-Nguyen,
  • Thanh-Nghi Do

摘要

Effective management of urban drainage networks is vital for resilience and infrastructure reliability. This paper introduces AquaNetCT, an automated framework for detecting, analyzing, and visualizing manholes in Can Tho city, Vietnam. The system employs computer vision for manhole detection, extracting geospatial and textual metadata based on OCR. Deep learning models (i.e., CNN-DenseNet121) classifies manhole types with over \(95\%\) F1-Score accuracy, enabling prioritized maintenance. The dataset extracted from images/frames is structured and stored into formats (i.e., JSON, CSV, \(\ldots \) ), and depicted using a graph-based algorithm designed to model the drainage network. This framework streamlines inspection workflows and enhances the operational efficiency and sustainability of smart city drainage management.