Scholarlink

One of the most important needs arising from urbanization is traffic management, in terms of reducing congestion and improving road safety. This paper introduces DeepTrafficFlow, a deep learning-based system for real-time traffic monitoring and congestion prediction. Processes video feeds through camera calibration, Region of Interest (ROI) selection, and video preprocessing. The system classifies vehicles like cars, bikes, and trucks with the help of CNN and Fast R-CNN, counts them, and uses this information in feeding into algorithms like logistic regression, decision trees, SVM, and random forests. The random forest model predicts the degree of traffic congestion so that an early intervention can be carried out with an accuracy of 98%. The system, powered by an NVIDIA GTX 1650 GPU, demonstrates robust performance with 96% precision in vehicle detection and tracking.

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

DeepTrafficFlow: A Deep Learning Approach for Real-Time Traffic Monitoring and Congestion Reduction in Smart Cities

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Vikas Kumar,
  • Shreyansh Tiwari,
  • Shraiyash Pandey,
  • Anurag Sinha,
  • Pronaya Bhattacharya,
  • Shweta Redkar,
  • Satyabrata Roy

摘要

One of the most important needs arising from urbanization is traffic management, in terms of reducing congestion and improving road safety. This paper introduces DeepTrafficFlow, a deep learning-based system for real-time traffic monitoring and congestion prediction. Processes video feeds through camera calibration, Region of Interest (ROI) selection, and video preprocessing. The system classifies vehicles like cars, bikes, and trucks with the help of CNN and Fast R-CNN, counts them, and uses this information in feeding into algorithms like logistic regression, decision trees, SVM, and random forests. The random forest model predicts the degree of traffic congestion so that an early intervention can be carried out with an accuracy of 98%. The system, powered by an NVIDIA GTX 1650 GPU, demonstrates robust performance with 96% precision in vehicle detection and tracking.